January 22, 2025 • 53 mins
Surviving bear markets and questionable life choices since 2022, this week's episode dives into validation woes, the magic of compound interest, and the unexpected joys of a hot bath and a good book.
Please support Game of Nodes and other great independent validators in the Cosmos. Make Needlecast, Kingnodes, RHINO, and Lavender.Five a part of your delegation strategy today!
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:00):
where it's managed, you go like, well, look at that in 10 years time and that could actually be
(00:05):
like compound interest is magic, right? It can work for you. Whereas like, if you just pay your
own salary in 10 years time, you're like, great, cool, I've paid my salary, but you're probably
better off just saying, you know, validation was fun, but you know, fucking maybe maybe I just go
work on my dance for the next 10 years and then welcome to Game of Modes, a weekly podcast from
(00:30):
independent value data teams. Welcome to Game of Modes, a weekly podcast from independent value
data teams surviving bear markets since 2022 still here. We put off last week's episode because of
illness on the podcast. They were shits all around. There was a lot. I was ready to go to be fair,
(00:52):
actually. I was all right. But everybody else was ill. So I had a hot bath and read Master and
Commander. So, oh, God, now you're reading it as well as watching it. I'm reading the series.
The whole series, like the whole the books, that exists. 20 bucks on eBay. That is a great
(01:15):
deal. That's a lot of content. How many books is that? 19. Although there's actually two more books.
There's actually two more books. There's a 20th, I think, and then he died while writing the 21st.
So there you go. It sounds like you need the bundle of holding in your life. Yeah. You ever
heard of this? What's that? The bundle of holding. The bundle of holding is like, it's like a humble
(01:39):
bundle where you pay like $5 for like 30 books. And sometimes it's, and it's by the authors. The
authors often do it, right? Sometimes you get stuff. I tend to get all of them just because I'm a
data hoarder. Yeah, I was going to say that that kind of feels like a hoarding thing. I remember
getting so excited the first time I saw that humble bundle. And then they became a burden to me.
(02:00):
And then they became a burden to me. I was like, I can just sleep with these. I've done it all these
games. I need to play them. Wait, explain the humble bundle. The humble bundle. Humble bundle.
Humble bundle is, it's originally for games where you pay like, you get to choose what you pay,
goes to charity, and you get like seven games. So you get to pay $5 for seven games. And often,
(02:24):
there's like good stuff in it, right? But it means that you end up ultimately having this massive
backlog of games that you're not actually interested in playing because you paid extra for
like just a couple of games that you wanted. Yeah, good model, right? Because you're like,
I hate games. And then, you know, the next time you have a long train journey, yes, I know.
Don't fucking say it. Next time on a long train journey, I'm just playing XCOM 2 again, aren't
(02:46):
I? Of course I am because I play XCOM 2. That's what I do on the train. When I'm like, Hey,
which version of lemmings are we going to play today? Do you actually play them? No, it just
seems like the type of thing that would be in a humble bundle. It's like, for a word. Yeah,
actually, hey, worms 2 is a fucking great game. That one stands up. Test of time.
(03:09):
That got me through lockdown. No, worms are no worms are we getting words? Well, part no,
not world party. Worms WMD. Oh, yeah, WMD. Playing that with the lads every Thursday.
That's great. What was that? Captain King. Is it Captain King? I don't know that one.
It's a real early game, like fucking back in the day. It's like before games were games.
(03:34):
This is Captain King. Before games were games. Real man.
So to our viewers who are expecting, this shit had to poke at you.
Really looking forward to our skip episode today, but he's not coming. So yeah.
(03:54):
And we had some confusion on the timings and dad had to go out for cigarettes.
So it's going to be next week. Yeah, there was a scheduling conflict due to some
calendar snafus. So typical gone shit, but that'll be here next week. So you'll be able to enjoy that.
Yeah, and we'll be as incisive as ever, I believe. So are we going to do a, I think we should do,
(04:18):
are we going to do an awards show? Because that was appetised, right? We said,
so the problem was for those listening, right? We, as professionals, we are, we're very professional,
obviously. We had a set of topics and also we have also done some previous years. So,
you know, posited on the group chat, some topics to the lads, and there was a deafening silence
(04:44):
of not getting a shit. Now, I appreciate that we have had what even fucking year are we in?
We have had nearly three years of bear market at this point.
You know what it is, the fray and the excitement maybe is not that I understand that, but no,
no, it's appointed by lack of energy. That's all. It's not that it's not that it's that the eco is
(05:05):
so fucking broad now that it's impossible to be on top of it all. And like, if we're out here trying
to like pick things, I mean, we're only picking things out of our own little world. And, you know,
I just don't have the energy to go and like educate myself on all of the other bullshit so that we
can be like, you know, some kind of accurate. If that makes any sense, I think we're just giving
(05:29):
awards to just random shit. I would agree with that. Yeah, it feels like as we've all expanded
more, there are less things that you can really pinpoint as award worthy, which makes it tough.
Which makes it tough. It's not as fun. Like our own sphere of the ecosystem is probably just a small
(05:50):
fraction. Like we're into the things we're into and we're not all into the same things. And like,
I might think something's great and you might never have fucking heard of it. So, you know,
that would be a fault of this podcast. I would say that we are failing with this podcast though.
Would it not? No, not real. This is sort of the forum. This is the forum. I mean, we're not fucking,
(06:10):
we're not a podcast of researchers. We don't sit there just like researching all day. So, we're
working, doing fucking work. And only sort of, you know, when you're working, you're only interested
in the things that make you money and are like, you know, business pertinent to business. You're
not on Twitter 25 hours a day fucking scrolling through shit making notes. We still make money.
(06:37):
Oh, yeah, there's money to be made. Yeah. Yeah, true. So I mean, I don't print it. I don't paint it.
Don't print it. Don't make it. Yeah, fair enough. It's been some time. It's been some time.
It's been some time. I think I think the problem I have more with the with the award show is it's
(06:57):
been quite some time really since all of our eggs as a company, at least in the crypto,
because we still obviously do other stuff weren't kind of in one basket. So it's kind of just like,
well, you know, there's only we were we were we were always doing fewer chains. And for the longest
time, only one's making us any money. So it's kind of like, well, we're interested in other things.
(07:19):
But, you know, DVT stuff, cool. Does it make us any money? No, you know, it just costs a server
costs every month, you know. So there's a whole bunch of things like that, I think, where it's
like things you're interested in. But like, you can't not sure you can slap an award on something
that's that's not a good business decision. Although that would eliminate a lot of things
(07:39):
in this case. That's not a business decision. You can't give that an award. Okay, well,
what are we going to do then? Fair point. To against myself there on DVT, though,
on DVT, at least it's kind of interesting. It is kind of interesting.
Yeah. Yes, you're right. It doesn't make any money, but it is kind of interesting.
(08:02):
It does make more of a public service.
More of a public service, which is funny, because, you know, that would be a good argument
if you were working in an area where some people weren't quite literally printing money.
So I don't know, like, it's yeah, I remember the other day, friend of the podcast hit us up
in the group and it's like, you guys are going to be keeping up your camera. If it was light
(08:26):
or SSV cluster or, you know, nodes, and we're like, yeah, man, it was SSV. It was like, yeah, man,
we're committed to keeping those things running. And then later on he messaged and he's like,
oh yeah, so I've put up a validator on yours guys nodes. And I'm like, cool. And he's like,
yeah, I've bought enough SSV to last seven years or something. And I'm like,
(08:48):
I don't think any of us even charge fees.
I, how do you charge?
How do you even charge fees?
You set the amount of SSV that you want to charge when you start to set the thing up.
And I, the thing I just put zero one or something like that.
Because I am paying the fees for it. So I'm pretty sure.
(09:09):
Okay. So there is some sort of fee. How much are you paying? Is it tell you?
It's one SSV per node per year.
Okay. So we all must have put in one SSV. It must have required something.
You can put zero. You can put zero, but everyone put in one.
Geez, I must be flush with SSV now.
I don't think I have any SS. Oh, well, I guess we will have some SSV.
(09:31):
Don't know how to claim it.
I think it's periodically paid out. I'm not sure though.
Oh, I can know this man. That was, that was one of those ones where it was set up using
quite a bit of automation when we were having a sleep regression. And, you know, anything that
could have been done, anything that was done where it is like the servers and it has alerting
in front of it and blah, blah, blah. That was all done with automated deployment stuff.
(09:54):
So I'm happy and cool that that all works. And it's got dashboards and whatnot.
But everything that was done via that SSV UI, which is also not great,
it was just a case of like, put in the numbers.
But I think it was like literally me in a chat with Shorts, who'd be like,
(10:16):
what are all these numbers mean? Shorts being like, oh, well, yes, here's how you do it.
And I'm just like, let's imagine that I am tired and really stupid right now.
What number do I put in box B is like, I literally don't put in five.
And I'm like, I've put in five and hit enter. Shorts is like, no,
no, that's the one number you couldn't put in. That's the last number you should have.
(10:42):
You always put, if you don't put one, you always put seven because seven is lucky
and it makes the blockchain go faster. And I'm sorry, I broke the whole blockchain.
I literally just had that conversation with Zenrock like three days ago,
where we hop on a call and I'm like, okay, imagine I've never used a terminal before
and I have a newborn and I'm really tired and I have like eight. Now let's continue and try
(11:06):
and figure this out. And the guy was like, sweet. And then it was like, I swear to God,
it was the single smoothest like paired programming session I've ever had in my life.
The paired programming, but like paired working through something. My God, it was glorious.
Treating me like an idiot and everything is deeply.
That's not a fantastic thing.
Zenrock was particularly confusing in some respects, maybe just because it's different.
(11:28):
Irrequiring Kubernetes definitely didn't help.
And how did you use Kubernetes?
Is this recent history we're talking about or a while ago?
No, this was literally like three days ago.
Oh, there's a binary is now.
We have binaries. Well, so you don't have to use helm anymore.
Yeah. So this is for their, their distributed signing, not for the validator.
(11:51):
Oh, right. Oh, fucking.
Yeah. So is this like their version of Horcrux or something?
Kind of, but not really. It'd be more like their version of SSV where it's all separate people
signing keys effectively. But it's not, again, it's not for the validator.
It's for kind of like a different side of their product.
(12:15):
Interesting.
Yeah. I was like, I stically note that on their website, they use the word decentralized multiple
times and they don't sign sources and even define the word decentralized.
My goodness. I mean, what does that mean?
I wonder if somebody was writing a paper on what that was, I mean,
lots of people are writing words on all that writing words and what that paper means,
(12:37):
writing papers and what that word means.
But I don't even know what you're saying then.
Fuck if I do, mate.
So last night, you can look forward to this because you go get through a whole bunch of
sleep regressions and then you can do potty training and you're going to sound as fucking
incoherent as I do.
We are, we are just, just got over a sleep regression and I'm still currently sleeping
(13:02):
on the floor of the office during the night. And my partner comes and swaps with me at 3am
and I come and take the morning shift when your man fucking starts to wake up.
So last night, Kelly said that the baby was great last night, slept most of the night,
(13:24):
you know, good night. And meanwhile, on the floor of the office, I fielded about fucking 50
page page duty fucking alarms. She came at 3am and I'm like, fuck off.
I am.
Spellings were real. That's a lot of fucking page duty action.
(13:45):
There was this fuck of a network that just kept stalling.
Couldn't be neutron because it just stalled once for 24 hours or 28 hours.
I kept fixing it and it kept stalling.
Oh, it was neutron. How did you fix it?
Oh, no, no, no, no, no, sorry. It wasn't, it wasn't neutron. It was a, it was an archive.
(14:06):
It's particularly a pain in the butt.
It's still unclear why, but I've been trying to figure it out with the team.
But anyway, it did just keep stalling. So now I've just put my super special little script
in there to make it not do that anymore by way of just restarting it when it stalls.
(14:28):
So that's the temporary fix at the moment.
When in doubt.
Yeah, I've actually got that script now as like, it goes on all of the servers and just
takes arguments. So it's like, you know, this just instead of having to like copy it across
and put it into, you know, run it in, I don't know, the background or whatever,
you can just, you know, run it straight out of the user bin on every server.
(14:54):
We rewrite it in Rust.
Networks got shit that stalls apparently.
Rewrite it in Rust, put it in a Docker container, call it production ready.
Open source it.
You can put bash scripts into user bin and just use this.
I did that. I actually have something very similar that I did use Docker for.
Docker is not happy when you try and restart a service outside of it. I'll have you know.
(15:17):
In fact, it's very unhappy about that.
Now if you think about it more than a second and why?
Yeah.
I Docker makes it easy to reproduce.
Yeah, basically.
But I didn't think about the fact that Docker, the kind of the purpose of it is to
not interfere with the services outside of it.
And so I was debugging it for like, probably a day like, what the hell?
I know this works. I know it works. I just know it works. It doesn't work.
(15:40):
And so I just have to run it natively now.
Ah, yes. So my new friend is just having in Ansible a folder filled with fucking
random scripts that I use and then just putting them into the user bin.
And then just using them. It's great.
(16:01):
Handy handy little helpers there.
Maybe I'll call that the role handy little helpers.
Handy little helpers.
The user bin handy little helpers.
Hey, just like I said, just put all those handy little helpers into a rust binary.
Open source it handy little helpers.rs.
Sell consultancy for your handy little helper service.
(16:21):
Give away the secrets, man.
The secret sources system.
We could totally restart.
We all know that.
Yeah.
Secret.
Shelling out.
I remember very early in my career as in like the first time I was really doing any ops work
over 10 years ago.
(16:42):
I was trying to basically get these reproducible things working on some servers.
We're moving to Docker.
This was when Docker was pre version one.
It was really, really early days.
And before Docker compose, before any of that, when Docker compose was in like
Alfred, it was called fig.
That's how long ago this was.
(17:02):
And we were running Ruby on Rails apps and needed to deploy them with a bunch of specific
to Ruby stuff.
And so I wrote this kind of Ruby wrapper for essentially like sort of Docker deployments
remotely that was a little bit inspired by Capistrano.
But because of how there wasn't any good way of actually ultimately doing the final bits
(17:28):
of orchestration that you needed other than just via the command line.
In the end, like because you could obviously, there are ways of interacting with the command
line via Ruby that use libraries that go like, oh, I'll just call this thing for you or whatever.
But then by some use, you're writing abstract object classes to like add additional arguments
to a thing because it takes more arguments than the developers have thought you'd need
(17:49):
or whatever.
You just go like system M out.
And so like it was this like, like the command line interface for this Ruby utility was like
actually all right.
But then when you dug right way down into the code, eventually it was just like
most of the harder Docker stuff was just system out.
Like it could have been done by bash by bash group.
(18:11):
And it was just like one of those like, oh, this is cool.
We can use this for this and that and the other.
And you're just like, just do not look in the internals.
Just do not look in the internals.
You realize I've just written a bash script, but in Ruby.
Yeah.
I was so proud of that for a little while because I was like, oh, well, it kind of works.
I thought about it a bit more.
I said, this is fucking stupid.
But yeah, I quite love bash scripting.
(18:33):
I think it's great.
It's great.
It's pretty hot take.
Yeah.
Like there are so many edge cases I can think of, but I couldn't describe like effectively over
a video where like you think you're doing it right.
And then one small syntactical problem ruins it and your day is effectively ruined.
I have never spent so long trying to understand something than reading the fucking bash manual.
(18:59):
It's fucking.
Yeah.
It's a really long B. There is so much shit in there and C. It just does all this stuff that's weird.
But I don't know when you get it, when you get it humming, it's humming.
But if your bash script is longer than about 20 lines and it's not a build script,
build scripts are the exception.
(19:20):
We allow build scripts, right?
Because they get run every build.
So if they break, we fix them very quickly.
It's not a big deal if you have a big, scary, silly build script.
People freak out about that too much.
But any other situation, your bash script is longer than like.
I don't want to say 30 lines.
I think 30 lines is the length of bash script where I look at it and I go,
(19:41):
that's going to cause problems.
Something like that.
It's about basically like one page without tabbing down in the terminal.
If you just less it and it's about a page, you're like, okay, okay, that's fine.
I'm all right with that.
And if it has like the little arrow down, I'm like, no, no, no.
(20:04):
I think this might be doing too much.
And I start to get that little tingling in the palms of your hands where you know something bad
is going to happen at some point.
So I do have a script that is exactly 99 lines long.
So yeah, I get over the 20.
I think I've got one that's like 400.
Yeah, definitely have longer ones.
This is just the first one I clicked on.
(20:25):
Eek.
In my defense, a lot of it is logging.
Yeah.
I mean, it would have to be 300 lines of logging for you to even get to null the size.
It might actually be for every function, I have the massive manual block of like six lines of
this is what's doing in this function.
(20:46):
So,
Ah, fair enough.
Oh, like actual code comments.
Code.
Yeah.
Yeah.
Yeah.
Okay.
The, the Juno build scripts were probably about 200 lines long back in the day.
But then I guess that was, that was building, building valid genesis file,
building the chain, launching the chain, uploading some smart contracts, making some accounts,
(21:07):
making everything work and then going like, does it all work now?
So that was a kind of, yeah, I would say if you just looked at that, like appropriate nothing,
you'd probably be like, this is fucking insanity.
But then at the same time, like, I think Reese wasn't scared by it when he looked at it.
I don't know what that says about the script.
But, you know,
yeah, maybe tentatively roll back my harshness about the, about the length of the bash scripts.
(21:31):
We'll see.
I do quite enjoy in bash that like most of the exits from blocks, like a case block,
the exit is ESAC.
And if it's like five, for example, just keeps going.
Yeah.
Anyway, it's a lot of mine.
Okey-doke.
Okey-doke.
(21:52):
Moving on.
Moving on.
That's a good, I'm glad that you enjoy your ESACs.
You've got to have something to keep your spirits up in these trying times.
Yeah.
Yeah.
It's the little things that, you know, really make life work.
Like, case backwards.
In fact.
Yeah.
(22:12):
ESAC on the fire.
Shit.
It's like, so it's, oh, hang on.
We've got, we've got a thing for this.
It's, it's pay your taxes time here.
Like, have you reported them and pay them and then you need to pay them for companies,
(22:34):
which means that it's the time of year where I look at our five year financials and remember
just how much, just how much we lost by staking in Cosmos.
And so that's fun as well.
So I think in the chat earlier, Rahm was like, oh, we're still in a bad way.
(22:55):
And it's like, well, you know, it's all relative, right?
Remember, June, June, I was $40 once.
And so, you know, if you, if you stake that June here.
I think, I think gauging overall market sentiment on the price of June is a little bit misplaced.
Well, no, you should probably gauge it on fucking Bitcoin, right?
(23:17):
Yeah. Well, like, you know, anything else, I think, like just a random main point on Salon
are probably better.
Yeah, I mean, I think that's probably true.
Probably like, you know, if you, if I think if you look at like the top 20, like,
you can say, if you look at maybe like Ethereum, Bitcoin, let's say, because they kind of are
in a class of their own in terms of what they do, although correlated, obviously.
(23:39):
And then, you know, some, do you know what they're, and obviously, as soon as I think
about this, they're obviously like a thousand, there'll be a thousand research papers on this,
on how you actually classify what the market's doing in terms of altcoins and also Bitcoin
and Ethereum. I could just look it up. Maybe I will. But I guess, like from our point of view,
the thing that's relevant, things most relevant to the majority of validators, given that it's
(24:04):
purely like, what's it called? Ethereum is just like a margins game.
The altcoins are always going to be where the future of our businesses is validators are, right?
Because they're the only ones that can have an upside to taking on counter-party risk,
essentially. Because the other, the parties are over in the other places.
(24:30):
But yeah.
The parties are over?
Well, it's like trying to elbow into Solana right now, right? You can jump in and pay your
and pay your two Solana, 15 Solana a month just to run your node. But now it's $250 per Solana,
or whatever it is, versus when the start is like 30 cents each, right?
Yeah. Sounds like that needs an adjustment. Is there a lot of Solana validators?
(24:55):
Hundreds, many hundreds, maybe even thousands, many, many, many.
All paying, what did you say, 15 Solana a month?
About 15 a month. Yeah.
So like three grand?
Yeah. You have to pay in order to like validate transactions because in Solana,
a transaction is how you sign. And so for every 0.2 second block, you're doing a transaction.
(25:19):
And every other validator is also doing transaction. It's one of the ways that they
kind of like game the system for there, how much transactions they're getting through
on the network. Because if there's a thousand validators and they're doing a transaction,
then suddenly they're saying they're doing, you know, 5,000 transactions per second.
When all that is doing is moving a block one forward.
Yeah. I remember actually that's similar to say when you looked at the stats on say when they first
(25:47):
were launched, there was like, you know, 40 transactions per block, which is, you know,
120 transactions per second, which was just like the oracle.
Yeah. Yeah. Yeah. It's like
the
120 million blocks. That's a lot of transactions.
(26:09):
Yeah. But the point dance getting in on those established networks is
the knowledge point, right?
Yeah. Yeah. So and, you know, even with like these Cosmos SDK or other altcoin networks,
after the initial launch, it is really hard to break into anyway, even just weeks or a month
after, because most of the initial staking is done in that period. And then it's pretty hard to move
(26:34):
stake from people after they do that. A lot of people just set it and forget it, collect their
rewards or whatever. So I mean, a great example of that is the validator, the so-called skips
validator on the hub right now. They rebranded, I remember what they were before.
Oh, yeah. Something 3S in 3S or something like that. Something like that. Yeah. And they changed
(26:55):
to skip whenever skip rebranded to interchain labs. Interchain ink. Sorry. Yeah. Yeah. Interchain
it. Interchain ink. Yeah. Well, regardless, validator rebranded. Ask next week. Yeah.
Yeah. Validator rebranded and changed the commission to 100%. And very few people have
left and they have 200,000 atoms staked to them. Yeah. I think a few over the years,
(27:20):
like a few people have pulled that move, haven't they? There was that,
um, fuck who was, I can't remember the name of them. They did it and they're like, oh,
it was a mistake. And then we're like, oh, we'll reduce it at like 1% a day for the next 100 days
or something. Yeah. I'm surprised they could ramp it up that quickly because you have like a
maximum day change. You have a maximum change, don't you? With the time? Yeah. The funny thing is,
(27:45):
is that the maximum change applies to going up, not down. Yeah. So they had the ability to change
it all at once. That's funny. I assumed it was bi-directional. So today I learned something
new about the Cosmos SDK. I think that's how it works anyway, from memory. I mean,
the more toxic thing is going up. So you would assume that that would be the way it was
implemented. But I guess I just assumed it was, it was both, but that could, that could need some
(28:10):
fact checking. I could need, I could need just a little bit of fact checking, but that's fine.
Yeah. Anyway, so skip next week. That'll be good. Yeah. There's quite a few things. Actually,
I think it's probably not so bad having them next week anyway, so we can make an appropriate
(28:32):
episode name, give a little bit of advertising, do a little bit more research, because they're into,
they're into quite a few things. And I'm wondering if they still have like, because
they've been into Chain, Inc. I guess the Cosmos stack is their primary, you know, product now
or objective. But I'm just wondering what happens also with all their other products that they have
(28:56):
throughout the Cosmos, which they had been working on previously. So be good to get
some more questions armed up around those types of projects and what that means for them.
Presumably they'll continue on with it. I'm curious how the acquisition worked. Like,
are they acquiring the entire team as an Aqua hire so that they can specifically work on the
(29:18):
Cosmos SDK? Or what? I doubt they're going to talk about it, but I just find those things really
interesting. Well, I noticed some emails are still skip emails and some are interchain, Inc. emails.
So yeah, I don't know. Maybe they've got like a skip team within interchain, Inc. now, like maybe
a division, the skip division. Who knows? We can ask. Speaking of Aqua hires, we also saw a
(29:46):
longstanding validator. That was news. So friends has been, I want to say acquired by
restake validator. I think I'd restake was big, big enough to acquire anybody ready.
(30:07):
That was sort of a surprise. So I don't know that I've ever heard of the actual restake
validator before. And looking at their list of networks that they support, maybe they're
reasonably new to the Cosmos. Right. So it's like sweet and a handful of others, right?
I think I saw some Solana. I'm not sure if I saw sweet and yeah, a grab bag of other
(30:37):
Cosmos SDK networks. But I guess they're making that was their way of making an entry easier
to the to the eco. Just have someone who's already got a few validators around. Yeah. And I see that
that Yuri has actually, no shit. I saw that Yuri has actually joined their team as well from
(31:00):
friends. So maybe like a merger type situation, not a hundred percent sure. Well, friends was
trying to sell like a year ago. I remember reading their pitch deck for why they're selling.
So I do imagine it was acquisition. I think they were trying to get out, right? Why would they
have a deck if they weren't trying to sell? So I bet it's some sort of like golden handcuff
(31:23):
type of thing where they want to maintain the connections for a year or two. And then the
founder of friends will bounce once that's over. How did you happen to cross that? Was it like,
like, did they did they contact you or? No, no. I don't remember. It was shared in like a ZIP group.
In a what group? A business development group. Yeah. Yeah. Interesting.
(31:52):
That was like, I'll buy that for a dollar. Yeah, I was just reading these comments. So
Rama says that skip is interchaining now. Skip doesn't exist anymore. All the investors were
paid out. Skip had skip now has 350 million dollars of interchain foundation funds.
Well, I just kind of raises more questions for me. Like, how were they bought? Were they bought
(32:15):
by the chain or they bought by the cosmos? I didn't vote on that, right? Like, what happened and how?
Skip now has the 350. Does that mean the skip the skip investors got paid out? 350 million?
Or that the new interchain ink is sitting on a treasury of 350 million?
(32:38):
I mean, I guess no offense, but like skip wasn't it would have been if somebody was
if somebody was paid out, it would have been at the the market rate of skip, which would have been
a couple of mill dollars, right? Max, because it's just it's just knowledge and people, right?
Yeah, sure. What I meant more was like, what what's the system for how that worked, right?
(32:59):
Like it wasn't bought using the community funds, right? Right on the actual chain.
Oh, I think I think Rama's answered you in the in the comments, which is that ICF is an independent
organization. Okay, so skip was bought by the ICF. Okay. Yes, ICF is, I think, well, it was a it was
(33:19):
a must. There must have been some money involved, but was also a merger because the original ICF
is now named something different after the merger because it was in IO and now it's interchain ink.
Well, it was interchain it was interchain GMBH because it was based in Germany, wasn't it?
Okay. And now it's a I only know that because I had to reference the
(33:42):
tendermint documentation repeatedly in one of my papers. And so I had to find out what
who the who the who I was referencing what what company and it was, I'm pretty certain it was a
GMBH company. Although that that's probably our date because that would have been last September
when I wrote those references. But I'm damned if I'm going back and reformatting my Bibtex. Oh, no,
(34:02):
no, no, that that file is done. Yeah, just because I can't last shit is is continually changing
because this that and the other doesn't mean my references are going to get messed with that
shit took ages. And they're done. So yeah, I definitely definitely remember the the pain of
referencing the Harvard referencing. It's not it is less bad if you use latex. But
(34:29):
but I didn't for my first draft, I just fucking bashed it out in one caffeinated psychopathic
effort, 20,000 words or whatever in Google Docs and then had to refine it down to 10.
And then reformatting latex, which was stupid. Just right. If okay, so here's right pro tip.
(34:49):
Anybody listening to the podcast, you're going to do any kind of academic work, even undergraduate,
to be perfectly honest, just just start by learning latex and and and just do it in latex
editor. Just don't don't bother wasting your time because you can just get preformatted
(35:10):
Bibtex files of every paper that's published in any kind of non Mickey Mouse journal or even ones
that are independently uploaded to archive. You can just get a Bibtex reference and stick it in
and move on with your life and overleaf is a good way to start. It's free. It's online. You can
access it via your institution. Don't be like this guy and write the whole fucking paper and then
(35:35):
have to refine it and then have to reformat it. It is demoralizing as fuck. Don't do that.
Little D says just use chat GBT to convert to latex.
That's a horrifying suggestion. Please don't do that.
Yes. So on on the subject of chat, GBT, recent subscriber right here, two thumbs up
(36:03):
and have also started using co pilot. And I don't know why I didn't jump on the train earlier.
It's quite handy that shit. It's glorious. It makes everything well, not everything easier.
I think that the real foot gun at this point is having the knowledge to know when things are wrong.
Yeah, that is definitely, definitely a point because it is quite frequently will feed you
(36:26):
some bullshit. But yeah, you know, very confident. I mean, that's the point, right?
Yeah, but it's good for scaffolding. So and it's good for small changes. So, you know, say,
for example, ginger to is quite confusing when you have a large ginger to template.
(36:48):
And say you want to like change or add in like a say you have a really long template and it's
it's got a bunch of loops in there to like bring things in depending on if they're defined and that
type of stuff. And you want to add in like another check, you could just select it all and tell it
to add in another check for this thing. And if you're like pretty precise with your prompt,
(37:13):
it'll do it and do it well. If you're not precise with your prompt, it may do something else.
I mean, it sounds like you just created something very bad, which maybe is the whole point of why
it's fine because it's fine. What do you mean created something bad? Explain yourself.
So you just described nested loops, I think, which, you know, generally pretty well,
(37:39):
ginger is a templating language is not aware of where ginger is now. I used to work in Python.
Yeah, but, you know, is there a different way to do it?
Well, you can obviously abstract out some logic so that you don't have to do nested loops and you
can also do things like list of comprehensions and intermediate functions. Like there are lots of
(38:01):
other ways of breaking out abstractions of things that need computing or checking in kind of because
a lot of what you would be doing is guard clauses. And actually, it sounds like you're
doing unnecessary iteration. And that sounds like exactly the sort of thing that you can probably
get chat GPD to come up with a solution for you. But it's going, it's back to Schultz's foot gun
(38:24):
thing. In fact, it's exactly I am just reiterating at longer length and more confused because I have
worked out a 10 hour day and then gone to a networking thing and then come on the podcast.
And I have no words left. I am just copying Schultz's homework of what he already said,
literally a minute ago. So I'll just say Schultz already said the thing about the foot gun and
(38:46):
the wrong and the okay. So, uh, Cope, oh, it's great for doing dumb shit then if that's what
you're into. I think it's just dependent on the environment you're in, I guess. Well, when you're
templating in Ansible, you only have so many fucking options. And you know, these, I don't see a way
around it. I'd love to have you take a look at it and you can tell me. I think Schultz had it.
(39:09):
Schultz had a point just there before I interrupted him. Well, I was going to go off on a bit of a
segue. So the nulls continue on this topic. I'm totally not doing it. Segue way out. Well, this
is irrelevant to the podcast, but I'm still kind of connected to my, um, my alma mater and chat GPT.
I think it's really damning for oncoming software engineers. Um, they won just the existence of
(39:36):
it means that junior engineers kind of having a hard time getting jobs, but two, they don't know
enough to know not to trust chat GPT and they can't kind of like intuit what's wrong with their answers,
like chat GPT's answers. Um, and I think that that's a serious, um, serious problem that
(39:56):
incoming software engineers have to deal with. Yeah. Fuck it. That's okay. Yeah. That's, I was
literally having a variant of that conversation earlier with, um, because not with the university
links, but I still do consulting and things in like trad tech stuff. And as somebody who's done
a lot of hiring before, I guess, like, especially like former colleagues and whatnot, they've just
(40:19):
gone on the traditional career and engineering career ladder while, you know, we were off being
DGNs over here. Um, but I guess I still, I still just about can, can like resuscitate
somebody else's opinion as my own to, to nod along with these conversations of people who have
responsible real jobs in tech. Um, but that is like something that has come up a fair bit is that
(40:40):
it's that, um, yeah, like that intuition, like you said, that you kind of, you just build up as
like work it like, um, uh, what's the term for journeymen, you know, like when you had craft
people in like olden times, you had like a novice apprentice and then you had a journeyman and
they would literally journey between different master crafts people and like learn from them
(41:01):
and build up their own intuition of how to create the outcome they were looking for, whether that
was a chair or a desk or a horse drawn carriage or whatever. And it was only when they became a, and,
and, and, you know, the, the idea of a master work, you know, it's sort of like, you cannot be
a master crafts person until you come up with your masterwork, right? And it's not possible
(41:24):
unless you go through all the pain of the iterative, iterative steps to build up the,
the reasoning of how to practically do these things. Because there's obviously book learning and,
and yada, yada, yada, but a lot of it is intuitive, especially in the past. And a lot of people have
made the connection between software engineering or programming, you know, programming as it is
outside of a big company, I suppose. And that kind of style of, of learning and progression. And
(41:49):
certainly at least for me, that's the, the progression that I went on from junior to senior
to whatever took that shape. And I saw it in people that I managed and mentored and whatever.
And I think it's quite a good model. And, and, and actually, yeah, and that's the thing I kind of
worry about, like, I suppose, with those sort of things, it's like similar to what you were
(42:10):
saying, Shorsi, is that like, if you, if you don't build that intuition early, then you,
you really don't have a foundation for much to build upon because you're always just building
feedback loops for intuitive processes that are usually below your conscious brain. And then the
bit that the conscious brain is doing is usually the feedback loops around, about learning and
(42:33):
problem solving, which are like how you, how you gather enough information for your brain to then
synthesize the solution. And that's quite a different problem space from the one that your
subconscious is working on. And the subconscious one only gets fed by experience. And chat, chat
gpt is basically short cutting that intuitive leap, it's short cutting that aha moment.
(42:58):
Yeah, exactly. I think that coding book came through an interesting example of that where I
know there's a lot of criticism coding boot camps, but some of the best engineers I've encountered
came from them. And I think that a coding boot camp will be a particularly bad spot for AI usage,
(43:20):
not to their fault, just that I've gone through them as like, kind of as an exercise to see what
it looks like. And a lot of them, you know, they do kind of advanced topics fairly quickly, and
they don't delve deep enough. And so it would be such an easy out to just be like, okay, well,
I'm not seeing this, but how do I, I don't know, convert HTML using CSS to make this widget bigger
(43:47):
as an example. Yeah, so like instead of instead of thinking harder on the on the thing, you just
get something else to do it. Well, but also sometimes it might genuinely be beyond you at that
point. So around the time I got my first job as a programmer, I mainly worked on front end stuff,
and I did some Ruby on Rails stuff. But you know, Ruby on Rails is relatively straightforward
(44:10):
within its own framework, like you can do a lot of back end stuff without having to know
how it all fits together. And I kind of knew I had this like big gap in my knowledge, right? I had
this very applied knowledge of like, decent HTML CSS JavaScript, okay, Ruby, okay. And then
I did one of those, I did one of those, you know, CS 101. There wasn't the MIT course where it might
(44:32):
have been like Udacity back in the day. I can't remember. And I remember doing like the basically
you work through an assignment, the whole course was gradually worked through an assignment,
where you essentially built a first a recreation of the original Google page rank algorithm.
That was what I was working up to. And it was done in Python. Come full circle to what would I
(44:57):
know about Python anyway. And I really, really distinctly remember like the thing that sticks
with me about that process is all the failures, right? Because and what was really interesting
about it is like you were saying shortly about when you get dumped into a topic area where it's
it's actually the gap between where you are and where the topic area is is potentially too wide.
I was working as a developer at that point. And I couldn't, there were some bits of that of that
(45:23):
algorithm, especially more complicated parts, which I actually couldn't couldn't figure out.
And if there had been chat GPT that I probably would have been able to get from a to b and solve it.
But I wouldn't have had any understanding. And what ended up happening was some of the final I
think I did get through it. And then I went on to the follow up and I literally got brickwalled by it.
(45:45):
And I couldn't do the next iteration of it. So I think I finished the CS 101, but I didn't
finish the 252, which was the follow up. And then many several years pass. And then I'm working on
a problem, which is a graph problem when I work for a data consultancy. And I realized I was working
on the problem that I had seen and failed at and never finished like several years before.
(46:08):
And obviously, in that context, I actually kind of by that point did know what I was doing,
solved it because it's part of my working day had to be done. And then just went back and looked it
up out of curiosity and was like, fuck me, no wonder I couldn't figure this out at the time.
But then had, you know, a very belated aha moment looking at it and going, oh, yeah, no,
actually, no, but I see, I see where my reasoning was going with what I've tried to do here. And
(46:32):
it was I was going completely the wrong direction. That's not how you solve this. Fine. But it's that,
you know, sometimes, sometimes learning something can be in the post for a very long time. And
maybe this is just me. I'm a very slow learner. So this is like very difficult of like, I don't
understand something. And then like years later, like fucking cosmos pre commits, man, that just
(46:57):
for the longest time, like, well, I was actually working on Juno Core just and then like years
later, you just like, oh, yeah, okay, mem pools. Everybody keeps talking about mem pools. And it's
not when you it's not until a long time later when you're like, but what are the essential
components of building a blockchain? And you're like, you build a blockchain in your head from
first principles, and you always end up with a mem pool. And you're like, you know, there's a lot
(47:19):
of things in life like that is what I'm saying. And if you can shortcut the, the reasoning required
to understand why you need a mem pool, you end up building a blockchain without a mem pool and
just be like, oh, fuck, we need a mem pool. Like, I don't know, that's probably a bad example.
Well, I get it though. It's like, you need to have that difficulty in order to overcome it.
(47:41):
And not being able to let things simmer because you get an instant
gratification answer is not doing yourself like you're doing yourself a disservice,
to a certain extent, like at this point, like, no, I'm not criticizing you directly in any way,
shape or form. I'm not saying that you shouldn't be using chat tbt.
This is like 25 minutes. This has been purely criticizing enough the last 25 minutes.
(48:02):
Yeah, exactly. All for all for using chat for co-pilot and chat tbt.
Now, for us, for us running a business and we need to use as like a periphery of what we're
actually doing, like, who cares? Right? Like, I'm like, we it's solving something.
The work that works. Exactly. Yeah. But for someone that's still building up their core
(48:24):
strengths in order to create their career, that's where I think it's it can be a damning tool.
Yeah, I mean, I'm not trying to make a career in fucking programming, but
I mean, I make a career in degeneracy, baby. Yeah, it suits our purpose. Even if it's doing
the wrong thing, even if I'm doing the wrong thing, it works perfectly for what I need it for.
(48:44):
So, you know, I don't have the benefit of being a programmer for the last fucking 20 years. So,
if I do things a little wrong, fuck it. I like it because it works.
But I use it as well. I have subscribed to it for like the last, I don't know, 12 months, 18
months. I use it mostly for like APIs. I don't like researching API. I'll just be like, Hey,
(49:07):
how do I send this over here? And then it'll be like, boom, I'll be like, cool.
I find it most useful for other people's API for shit documentation. God damn.
I find it most useful for like brainstorming ideas on how to solve a problem anyway. I use it a bit
in VS code for helping me just quickly change shit. But in like, you know, if I got like a more
(49:33):
complex problem with what I'm trying to do, I'll just talk to it for like an hour and say, you
know, start off with my issue and then explore with it the options it thinks is the way to do
things and then maybe come up with a scaffold of how I want to attack that problem and then start
trying to attack it. But I find it quite handy for that type of stuff. Not so much co-pilot,
(49:56):
but definitely chat GPT. Yeah. So I mean,
So you basically pair programming with it?
Yeah, because, you know, I'm not a, I don't have the benefit of years of programming and I'm
just building. So it's quite handy for that. Thinking out a problem, I guess, thinking.
I've got a, out of pure curiosity, this will take longer, but put a little,
(50:23):
put a little mascot or a duck or something on your desk.
And ask the duck. I knew that's where that was coming. I was going to say,
this sounds just like rubber ducky debugging the game.
It would take longer, but see whether you get to the same solution
and how that, whether there's a difference in satisfaction.
I really can't talk today. Satisfaction. I'd be really, really interested because like,
(50:45):
I've got literally got on my desk here, I talked to Totoro and in Totoro's little bag,
I also have the Buddha in case Totoro and cutting it. And you need to,
you need to appeal to a higher, a higher authority. You ask him, you ask the Buddha,
and if you need the highest authority of all, you got president business,
(51:06):
you got president business in case you really need with a little light.
This is what my Tanooki's for.
Exactly. You're just like, president business, could you help me? I can't solve this thing.
Okay. So what is asking the thing going to get for me? Explain.
Oh, right. So the point is you're always,
the solution is always going to come from your hind brain, right? It's going to come from an
(51:29):
intuitive leap. So what you're always trying to do is feed data to your brain to synthesize the
problem and then make the intuitive leap to solve it. So by ideating out loud, you iterate towards
a solution even in the absence of somebody talking back to you. So my thesis here, right,
is that chat GPT is useful, but the majority of the usefulness is your own brain chewing the problem.
(51:55):
And I'm just, I'm just naively curious, I suppose, whether you could get the same result
from describing your problem out loud to an inanimate object and letting your subconscious
figure it out. Because that's what junior developers were taught to do when, when I think
in some instances, yes, and other instances, no, because I don't necessarily have all of the inputs
(52:19):
required to be able to make the intuitive leap. For example, like there might be a technology
or a software that I'm just not aware of and wouldn't be able to arrive at on my own, a method
even that's employed by a particular software or package or something that I don't know about.
(52:40):
I might need to do something, but not think that it's possible with what is already sloshing around
just because I don't have the breadth of knowledge already to arrive at that. But if it's just a logic
problem or something like that within a program, probably, yeah, I would assume that type of stuff
(53:00):
I'd probably just think on anyway, while I'm like taking a shit. So I'd ask the wall, I'd ask the
Dunny.
Always ask the Dunny if you're stuck.
Yes, shits quite often take an hour and a half instead of 10 minutes just because I'm staring
at the corner of the shitter thinking.
And we too could have this problem if you decide to take the exciting career choice of being a
(53:24):
validator.
where it's managed, you go like, well, look at that in 10 years time and that could actually be
(00:05):
like compound interest is magic, right? It can work for you. Whereas like, if you just pay your
own salary in 10 years time, you're like, great, cool, I've paid my salary, but you're probably
better off just saying, you know, validation was fun, but you know, fucking maybe maybe I just go
work on my dance for the next 10 years and then welcome to Game of Modes, a weekly podcast from
(00:30):
independent value data teams. Welcome to Game of Modes, a weekly podcast from independent value
data teams surviving bear markets since 2022 still here. We put off last week's episode because of
illness on the podcast. They were shits all around. There was a lot. I was ready to go to be fair,
(00:52):
actually. I was all right. But everybody else was ill. So I had a hot bath and read Master and
Commander. So, oh, God, now you're reading it as well as watching it. I'm reading the series.
The whole series, like the whole the books, that exists. 20 bucks on eBay. That is a great
(01:15):
deal. That's a lot of content. How many books is that? 19. Although there's actually two more books.
There's actually two more books. There's a 20th, I think, and then he died while writing the 21st.
So there you go. It sounds like you need the bundle of holding in your life. Yeah. You ever
heard of this? What's that? The bundle of holding. The bundle of holding is like, it's like a humble
(01:39):
bundle where you pay like $5 for like 30 books. And sometimes it's, and it's by the authors. The
authors often do it, right? Sometimes you get stuff. I tend to get all of them just because I'm a
data hoarder. Yeah, I was going to say that that kind of feels like a hoarding thing. I remember
getting so excited the first time I saw that humble bundle. And then they became a burden to me.
(02:00):
And then they became a burden to me. I was like, I can just sleep with these. I've done it all these
games. I need to play them. Wait, explain the humble bundle. The humble bundle. Humble bundle.
Humble bundle is, it's originally for games where you pay like, you get to choose what you pay,
goes to charity, and you get like seven games. So you get to pay $5 for seven games. And often,
(02:24):
there's like good stuff in it, right? But it means that you end up ultimately having this massive
backlog of games that you're not actually interested in playing because you paid extra for
like just a couple of games that you wanted. Yeah, good model, right? Because you're like,
I hate games. And then, you know, the next time you have a long train journey, yes, I know.
Don't fucking say it. Next time on a long train journey, I'm just playing XCOM 2 again, aren't
(02:46):
I? Of course I am because I play XCOM 2. That's what I do on the train. When I'm like, Hey,
which version of lemmings are we going to play today? Do you actually play them? No, it just
seems like the type of thing that would be in a humble bundle. It's like, for a word. Yeah,
actually, hey, worms 2 is a fucking great game. That one stands up. Test of time.
(03:09):
That got me through lockdown. No, worms are no worms are we getting words? Well, part no,
not world party. Worms WMD. Oh, yeah, WMD. Playing that with the lads every Thursday.
That's great. What was that? Captain King. Is it Captain King? I don't know that one.
It's a real early game, like fucking back in the day. It's like before games were games.
(03:34):
This is Captain King. Before games were games. Real man.
So to our viewers who are expecting, this shit had to poke at you.
Really looking forward to our skip episode today, but he's not coming. So yeah.
(03:54):
And we had some confusion on the timings and dad had to go out for cigarettes.
So it's going to be next week. Yeah, there was a scheduling conflict due to some
calendar snafus. So typical gone shit, but that'll be here next week. So you'll be able to enjoy that.
Yeah, and we'll be as incisive as ever, I believe. So are we going to do a, I think we should do,
(04:18):
are we going to do an awards show? Because that was appetised, right? We said,
so the problem was for those listening, right? We, as professionals, we are, we're very professional,
obviously. We had a set of topics and also we have also done some previous years. So,
you know, posited on the group chat, some topics to the lads, and there was a deafening silence
(04:44):
of not getting a shit. Now, I appreciate that we have had what even fucking year are we in?
We have had nearly three years of bear market at this point.
You know what it is, the fray and the excitement maybe is not that I understand that, but no,
no, it's appointed by lack of energy. That's all. It's not that it's not that it's that the eco is
(05:05):
so fucking broad now that it's impossible to be on top of it all. And like, if we're out here trying
to like pick things, I mean, we're only picking things out of our own little world. And, you know,
I just don't have the energy to go and like educate myself on all of the other bullshit so that we
can be like, you know, some kind of accurate. If that makes any sense, I think we're just giving
(05:29):
awards to just random shit. I would agree with that. Yeah, it feels like as we've all expanded
more, there are less things that you can really pinpoint as award worthy, which makes it tough.
Which makes it tough. It's not as fun. Like our own sphere of the ecosystem is probably just a small
(05:50):
fraction. Like we're into the things we're into and we're not all into the same things. And like,
I might think something's great and you might never have fucking heard of it. So, you know,
that would be a fault of this podcast. I would say that we are failing with this podcast though.
Would it not? No, not real. This is sort of the forum. This is the forum. I mean, we're not fucking,
(06:10):
we're not a podcast of researchers. We don't sit there just like researching all day. So, we're
working, doing fucking work. And only sort of, you know, when you're working, you're only interested
in the things that make you money and are like, you know, business pertinent to business. You're
not on Twitter 25 hours a day fucking scrolling through shit making notes. We still make money.
(06:37):
Oh, yeah, there's money to be made. Yeah. Yeah, true. So I mean, I don't print it. I don't paint it.
Don't print it. Don't make it. Yeah, fair enough. It's been some time. It's been some time.
It's been some time. I think I think the problem I have more with the with the award show is it's
(06:57):
been quite some time really since all of our eggs as a company, at least in the crypto,
because we still obviously do other stuff weren't kind of in one basket. So it's kind of just like,
well, you know, there's only we were we were we were always doing fewer chains. And for the longest
time, only one's making us any money. So it's kind of like, well, we're interested in other things.
(07:19):
But, you know, DVT stuff, cool. Does it make us any money? No, you know, it just costs a server
costs every month, you know. So there's a whole bunch of things like that, I think, where it's
like things you're interested in. But like, you can't not sure you can slap an award on something
that's that's not a good business decision. Although that would eliminate a lot of things
(07:39):
in this case. That's not a business decision. You can't give that an award. Okay, well,
what are we going to do then? Fair point. To against myself there on DVT, though,
on DVT, at least it's kind of interesting. It is kind of interesting.
Yeah. Yes, you're right. It doesn't make any money, but it is kind of interesting.
(08:02):
It does make more of a public service.
More of a public service, which is funny, because, you know, that would be a good argument
if you were working in an area where some people weren't quite literally printing money.
So I don't know, like, it's yeah, I remember the other day, friend of the podcast hit us up
in the group and it's like, you guys are going to be keeping up your camera. If it was light
(08:26):
or SSV cluster or, you know, nodes, and we're like, yeah, man, it was SSV. It was like, yeah, man,
we're committed to keeping those things running. And then later on he messaged and he's like,
oh yeah, so I've put up a validator on yours guys nodes. And I'm like, cool. And he's like,
yeah, I've bought enough SSV to last seven years or something. And I'm like,
(08:48):
I don't think any of us even charge fees.
I, how do you charge?
How do you even charge fees?
You set the amount of SSV that you want to charge when you start to set the thing up.
And I, the thing I just put zero one or something like that.
Because I am paying the fees for it. So I'm pretty sure.
(09:09):
Okay. So there is some sort of fee. How much are you paying? Is it tell you?
It's one SSV per node per year.
Okay. So we all must have put in one SSV. It must have required something.
You can put zero. You can put zero, but everyone put in one.
Geez, I must be flush with SSV now.
I don't think I have any SS. Oh, well, I guess we will have some SSV.
(09:31):
Don't know how to claim it.
I think it's periodically paid out. I'm not sure though.
Oh, I can know this man. That was, that was one of those ones where it was set up using
quite a bit of automation when we were having a sleep regression. And, you know, anything that
could have been done, anything that was done where it is like the servers and it has alerting
in front of it and blah, blah, blah. That was all done with automated deployment stuff.
(09:54):
So I'm happy and cool that that all works. And it's got dashboards and whatnot.
But everything that was done via that SSV UI, which is also not great,
it was just a case of like, put in the numbers.
But I think it was like literally me in a chat with Shorts, who'd be like,
(10:16):
what are all these numbers mean? Shorts being like, oh, well, yes, here's how you do it.
And I'm just like, let's imagine that I am tired and really stupid right now.
What number do I put in box B is like, I literally don't put in five.
And I'm like, I've put in five and hit enter. Shorts is like, no,
no, that's the one number you couldn't put in. That's the last number you should have.
(10:42):
You always put, if you don't put one, you always put seven because seven is lucky
and it makes the blockchain go faster. And I'm sorry, I broke the whole blockchain.
I literally just had that conversation with Zenrock like three days ago,
where we hop on a call and I'm like, okay, imagine I've never used a terminal before
and I have a newborn and I'm really tired and I have like eight. Now let's continue and try
(11:06):
and figure this out. And the guy was like, sweet. And then it was like, I swear to God,
it was the single smoothest like paired programming session I've ever had in my life.
The paired programming, but like paired working through something. My God, it was glorious.
Treating me like an idiot and everything is deeply.
That's not a fantastic thing.
Zenrock was particularly confusing in some respects, maybe just because it's different.
(11:28):
Irrequiring Kubernetes definitely didn't help.
And how did you use Kubernetes?
Is this recent history we're talking about or a while ago?
No, this was literally like three days ago.
Oh, there's a binary is now.
We have binaries. Well, so you don't have to use helm anymore.
Yeah. So this is for their, their distributed signing, not for the validator.
(11:51):
Oh, right. Oh, fucking.
Yeah. So is this like their version of Horcrux or something?
Kind of, but not really. It'd be more like their version of SSV where it's all separate people
signing keys effectively. But it's not, again, it's not for the validator.
It's for kind of like a different side of their product.
(12:15):
Interesting.
Yeah. I was like, I stically note that on their website, they use the word decentralized multiple
times and they don't sign sources and even define the word decentralized.
My goodness. I mean, what does that mean?
I wonder if somebody was writing a paper on what that was, I mean,
lots of people are writing words on all that writing words and what that paper means,
(12:37):
writing papers and what that word means.
But I don't even know what you're saying then.
Fuck if I do, mate.
So last night, you can look forward to this because you go get through a whole bunch of
sleep regressions and then you can do potty training and you're going to sound as fucking
incoherent as I do.
We are, we are just, just got over a sleep regression and I'm still currently sleeping
(13:02):
on the floor of the office during the night. And my partner comes and swaps with me at 3am
and I come and take the morning shift when your man fucking starts to wake up.
So last night, Kelly said that the baby was great last night, slept most of the night,
(13:24):
you know, good night. And meanwhile, on the floor of the office, I fielded about fucking 50
page page duty fucking alarms. She came at 3am and I'm like, fuck off.
I am.
Spellings were real. That's a lot of fucking page duty action.
(13:45):
There was this fuck of a network that just kept stalling.
Couldn't be neutron because it just stalled once for 24 hours or 28 hours.
I kept fixing it and it kept stalling.
Oh, it was neutron. How did you fix it?
Oh, no, no, no, no, no, sorry. It wasn't, it wasn't neutron. It was a, it was an archive.
(14:06):
It's particularly a pain in the butt.
It's still unclear why, but I've been trying to figure it out with the team.
But anyway, it did just keep stalling. So now I've just put my super special little script
in there to make it not do that anymore by way of just restarting it when it stalls.
(14:28):
So that's the temporary fix at the moment.
When in doubt.
Yeah, I've actually got that script now as like, it goes on all of the servers and just
takes arguments. So it's like, you know, this just instead of having to like copy it across
and put it into, you know, run it in, I don't know, the background or whatever,
you can just, you know, run it straight out of the user bin on every server.
(14:54):
We rewrite it in Rust.
Networks got shit that stalls apparently.
Rewrite it in Rust, put it in a Docker container, call it production ready.
Open source it.
You can put bash scripts into user bin and just use this.
I did that. I actually have something very similar that I did use Docker for.
Docker is not happy when you try and restart a service outside of it. I'll have you know.
(15:17):
In fact, it's very unhappy about that.
Now if you think about it more than a second and why?
Yeah.
I Docker makes it easy to reproduce.
Yeah, basically.
But I didn't think about the fact that Docker, the kind of the purpose of it is to
not interfere with the services outside of it.
And so I was debugging it for like, probably a day like, what the hell?
I know this works. I know it works. I just know it works. It doesn't work.
(15:40):
And so I just have to run it natively now.
Ah, yes. So my new friend is just having in Ansible a folder filled with fucking
random scripts that I use and then just putting them into the user bin.
And then just using them. It's great.
(16:01):
Handy handy little helpers there.
Maybe I'll call that the role handy little helpers.
Handy little helpers.
The user bin handy little helpers.
Hey, just like I said, just put all those handy little helpers into a rust binary.
Open source it handy little helpers.rs.
Sell consultancy for your handy little helper service.
(16:21):
Give away the secrets, man.
The secret sources system.
We could totally restart.
We all know that.
Yeah.
Secret.
Shelling out.
I remember very early in my career as in like the first time I was really doing any ops work
over 10 years ago.
(16:42):
I was trying to basically get these reproducible things working on some servers.
We're moving to Docker.
This was when Docker was pre version one.
It was really, really early days.
And before Docker compose, before any of that, when Docker compose was in like
Alfred, it was called fig.
That's how long ago this was.
(17:02):
And we were running Ruby on Rails apps and needed to deploy them with a bunch of specific
to Ruby stuff.
And so I wrote this kind of Ruby wrapper for essentially like sort of Docker deployments
remotely that was a little bit inspired by Capistrano.
But because of how there wasn't any good way of actually ultimately doing the final bits
(17:28):
of orchestration that you needed other than just via the command line.
In the end, like because you could obviously, there are ways of interacting with the command
line via Ruby that use libraries that go like, oh, I'll just call this thing for you or whatever.
But then by some use, you're writing abstract object classes to like add additional arguments
to a thing because it takes more arguments than the developers have thought you'd need
(17:49):
or whatever.
You just go like system M out.
And so like it was this like, like the command line interface for this Ruby utility was like
actually all right.
But then when you dug right way down into the code, eventually it was just like
most of the harder Docker stuff was just system out.
Like it could have been done by bash by bash group.
(18:11):
And it was just like one of those like, oh, this is cool.
We can use this for this and that and the other.
And you're just like, just do not look in the internals.
Just do not look in the internals.
You realize I've just written a bash script, but in Ruby.
Yeah.
I was so proud of that for a little while because I was like, oh, well, it kind of works.
I thought about it a bit more.
I said, this is fucking stupid.
But yeah, I quite love bash scripting.
(18:33):
I think it's great.
It's great.
It's pretty hot take.
Yeah.
Like there are so many edge cases I can think of, but I couldn't describe like effectively over
a video where like you think you're doing it right.
And then one small syntactical problem ruins it and your day is effectively ruined.
I have never spent so long trying to understand something than reading the fucking bash manual.
(18:59):
It's fucking.
Yeah.
It's a really long B. There is so much shit in there and C. It just does all this stuff that's weird.
But I don't know when you get it, when you get it humming, it's humming.
But if your bash script is longer than about 20 lines and it's not a build script,
build scripts are the exception.
(19:20):
We allow build scripts, right?
Because they get run every build.
So if they break, we fix them very quickly.
It's not a big deal if you have a big, scary, silly build script.
People freak out about that too much.
But any other situation, your bash script is longer than like.
I don't want to say 30 lines.
I think 30 lines is the length of bash script where I look at it and I go,
(19:41):
that's going to cause problems.
Something like that.
It's about basically like one page without tabbing down in the terminal.
If you just less it and it's about a page, you're like, okay, okay, that's fine.
I'm all right with that.
And if it has like the little arrow down, I'm like, no, no, no.
(20:04):
I think this might be doing too much.
And I start to get that little tingling in the palms of your hands where you know something bad
is going to happen at some point.
So I do have a script that is exactly 99 lines long.
So yeah, I get over the 20.
I think I've got one that's like 400.
Yeah, definitely have longer ones.
This is just the first one I clicked on.
(20:25):
Eek.
In my defense, a lot of it is logging.
Yeah.
I mean, it would have to be 300 lines of logging for you to even get to null the size.
It might actually be for every function, I have the massive manual block of like six lines of
this is what's doing in this function.
(20:46):
So,
Ah, fair enough.
Oh, like actual code comments.
Code.
Yeah.
Yeah.
Yeah.
Okay.
The, the Juno build scripts were probably about 200 lines long back in the day.
But then I guess that was, that was building, building valid genesis file,
building the chain, launching the chain, uploading some smart contracts, making some accounts,
(21:07):
making everything work and then going like, does it all work now?
So that was a kind of, yeah, I would say if you just looked at that, like appropriate nothing,
you'd probably be like, this is fucking insanity.
But then at the same time, like, I think Reese wasn't scared by it when he looked at it.
I don't know what that says about the script.
But, you know,
yeah, maybe tentatively roll back my harshness about the, about the length of the bash scripts.
(21:31):
We'll see.
I do quite enjoy in bash that like most of the exits from blocks, like a case block,
the exit is ESAC.
And if it's like five, for example, just keeps going.
Yeah.
Anyway, it's a lot of mine.
Okey-doke.
Okey-doke.
(21:52):
Moving on.
Moving on.
That's a good, I'm glad that you enjoy your ESACs.
You've got to have something to keep your spirits up in these trying times.
Yeah.
Yeah.
It's the little things that, you know, really make life work.
Like, case backwards.
In fact.
Yeah.
(22:12):
ESAC on the fire.
Shit.
It's like, so it's, oh, hang on.
We've got, we've got a thing for this.
It's, it's pay your taxes time here.
Like, have you reported them and pay them and then you need to pay them for companies,
(22:34):
which means that it's the time of year where I look at our five year financials and remember
just how much, just how much we lost by staking in Cosmos.
And so that's fun as well.
So I think in the chat earlier, Rahm was like, oh, we're still in a bad way.
(22:55):
And it's like, well, you know, it's all relative, right?
Remember, June, June, I was $40 once.
And so, you know, if you, if you stake that June here.
I think, I think gauging overall market sentiment on the price of June is a little bit misplaced.
Well, no, you should probably gauge it on fucking Bitcoin, right?
(23:17):
Yeah. Well, like, you know, anything else, I think, like just a random main point on Salon
are probably better.
Yeah, I mean, I think that's probably true.
Probably like, you know, if you, if I think if you look at like the top 20, like,
you can say, if you look at maybe like Ethereum, Bitcoin, let's say, because they kind of are
in a class of their own in terms of what they do, although correlated, obviously.
(23:39):
And then, you know, some, do you know what they're, and obviously, as soon as I think
about this, they're obviously like a thousand, there'll be a thousand research papers on this,
on how you actually classify what the market's doing in terms of altcoins and also Bitcoin
and Ethereum. I could just look it up. Maybe I will. But I guess, like from our point of view,
the thing that's relevant, things most relevant to the majority of validators, given that it's
(24:04):
purely like, what's it called? Ethereum is just like a margins game.
The altcoins are always going to be where the future of our businesses is validators are, right?
Because they're the only ones that can have an upside to taking on counter-party risk,
essentially. Because the other, the parties are over in the other places.
(24:30):
But yeah.
The parties are over?
Well, it's like trying to elbow into Solana right now, right? You can jump in and pay your
and pay your two Solana, 15 Solana a month just to run your node. But now it's $250 per Solana,
or whatever it is, versus when the start is like 30 cents each, right?
Yeah. Sounds like that needs an adjustment. Is there a lot of Solana validators?
(24:55):
Hundreds, many hundreds, maybe even thousands, many, many, many.
All paying, what did you say, 15 Solana a month?
About 15 a month. Yeah.
So like three grand?
Yeah. You have to pay in order to like validate transactions because in Solana,
a transaction is how you sign. And so for every 0.2 second block, you're doing a transaction.
(25:19):
And every other validator is also doing transaction. It's one of the ways that they
kind of like game the system for there, how much transactions they're getting through
on the network. Because if there's a thousand validators and they're doing a transaction,
then suddenly they're saying they're doing, you know, 5,000 transactions per second.
When all that is doing is moving a block one forward.
Yeah. I remember actually that's similar to say when you looked at the stats on say when they first
(25:47):
were launched, there was like, you know, 40 transactions per block, which is, you know,
120 transactions per second, which was just like the oracle.
Yeah. Yeah. Yeah. It's like
the
120 million blocks. That's a lot of transactions.
(26:09):
Yeah. But the point dance getting in on those established networks is
the knowledge point, right?
Yeah. Yeah. So and, you know, even with like these Cosmos SDK or other altcoin networks,
after the initial launch, it is really hard to break into anyway, even just weeks or a month
after, because most of the initial staking is done in that period. And then it's pretty hard to move
(26:34):
stake from people after they do that. A lot of people just set it and forget it, collect their
rewards or whatever. So I mean, a great example of that is the validator, the so-called skips
validator on the hub right now. They rebranded, I remember what they were before.
Oh, yeah. Something 3S in 3S or something like that. Something like that. Yeah. And they changed
(26:55):
to skip whenever skip rebranded to interchain labs. Interchain ink. Sorry. Yeah. Yeah. Interchain
it. Interchain ink. Yeah. Well, regardless, validator rebranded. Ask next week. Yeah.
Yeah. Validator rebranded and changed the commission to 100%. And very few people have
left and they have 200,000 atoms staked to them. Yeah. I think a few over the years,
(27:20):
like a few people have pulled that move, haven't they? There was that,
um, fuck who was, I can't remember the name of them. They did it and they're like, oh,
it was a mistake. And then we're like, oh, we'll reduce it at like 1% a day for the next 100 days
or something. Yeah. I'm surprised they could ramp it up that quickly because you have like a
maximum day change. You have a maximum change, don't you? With the time? Yeah. The funny thing is,
(27:45):
is that the maximum change applies to going up, not down. Yeah. So they had the ability to change
it all at once. That's funny. I assumed it was bi-directional. So today I learned something
new about the Cosmos SDK. I think that's how it works anyway, from memory. I mean,
the more toxic thing is going up. So you would assume that that would be the way it was
implemented. But I guess I just assumed it was, it was both, but that could, that could need some
(28:10):
fact checking. I could need, I could need just a little bit of fact checking, but that's fine.
Yeah. Anyway, so skip next week. That'll be good. Yeah. There's quite a few things. Actually,
I think it's probably not so bad having them next week anyway, so we can make an appropriate
(28:32):
episode name, give a little bit of advertising, do a little bit more research, because they're into,
they're into quite a few things. And I'm wondering if they still have like, because
they've been into Chain, Inc. I guess the Cosmos stack is their primary, you know, product now
or objective. But I'm just wondering what happens also with all their other products that they have
(28:56):
throughout the Cosmos, which they had been working on previously. So be good to get
some more questions armed up around those types of projects and what that means for them.
Presumably they'll continue on with it. I'm curious how the acquisition worked. Like,
are they acquiring the entire team as an Aqua hire so that they can specifically work on the
(29:18):
Cosmos SDK? Or what? I doubt they're going to talk about it, but I just find those things really
interesting. Well, I noticed some emails are still skip emails and some are interchain, Inc. emails.
So yeah, I don't know. Maybe they've got like a skip team within interchain, Inc. now, like maybe
a division, the skip division. Who knows? We can ask. Speaking of Aqua hires, we also saw a
(29:46):
longstanding validator. That was news. So friends has been, I want to say acquired by
restake validator. I think I'd restake was big, big enough to acquire anybody ready.
(30:07):
That was sort of a surprise. So I don't know that I've ever heard of the actual restake
validator before. And looking at their list of networks that they support, maybe they're
reasonably new to the Cosmos. Right. So it's like sweet and a handful of others, right?
I think I saw some Solana. I'm not sure if I saw sweet and yeah, a grab bag of other
(30:37):
Cosmos SDK networks. But I guess they're making that was their way of making an entry easier
to the to the eco. Just have someone who's already got a few validators around. Yeah. And I see that
that Yuri has actually, no shit. I saw that Yuri has actually joined their team as well from
(31:00):
friends. So maybe like a merger type situation, not a hundred percent sure. Well, friends was
trying to sell like a year ago. I remember reading their pitch deck for why they're selling.
So I do imagine it was acquisition. I think they were trying to get out, right? Why would they
have a deck if they weren't trying to sell? So I bet it's some sort of like golden handcuff
(31:23):
type of thing where they want to maintain the connections for a year or two. And then the
founder of friends will bounce once that's over. How did you happen to cross that? Was it like,
like, did they did they contact you or? No, no. I don't remember. It was shared in like a ZIP group.
In a what group? A business development group. Yeah. Yeah. Interesting.
(31:52):
That was like, I'll buy that for a dollar. Yeah, I was just reading these comments. So
Rama says that skip is interchaining now. Skip doesn't exist anymore. All the investors were
paid out. Skip had skip now has 350 million dollars of interchain foundation funds.
Well, I just kind of raises more questions for me. Like, how were they bought? Were they bought
(32:15):
by the chain or they bought by the cosmos? I didn't vote on that, right? Like, what happened and how?
Skip now has the 350. Does that mean the skip the skip investors got paid out? 350 million?
Or that the new interchain ink is sitting on a treasury of 350 million?
(32:38):
I mean, I guess no offense, but like skip wasn't it would have been if somebody was
if somebody was paid out, it would have been at the the market rate of skip, which would have been
a couple of mill dollars, right? Max, because it's just it's just knowledge and people, right?
Yeah, sure. What I meant more was like, what what's the system for how that worked, right?
(32:59):
Like it wasn't bought using the community funds, right? Right on the actual chain.
Oh, I think I think Rama's answered you in the in the comments, which is that ICF is an independent
organization. Okay, so skip was bought by the ICF. Okay. Yes, ICF is, I think, well, it was a it was
(33:19):
a must. There must have been some money involved, but was also a merger because the original ICF
is now named something different after the merger because it was in IO and now it's interchain ink.
Well, it was interchain it was interchain GMBH because it was based in Germany, wasn't it?
Okay. And now it's a I only know that because I had to reference the
(33:42):
tendermint documentation repeatedly in one of my papers. And so I had to find out what
who the who the who I was referencing what what company and it was, I'm pretty certain it was a
GMBH company. Although that that's probably our date because that would have been last September
when I wrote those references. But I'm damned if I'm going back and reformatting my Bibtex. Oh, no,
(34:02):
no, no, that that file is done. Yeah, just because I can't last shit is is continually changing
because this that and the other doesn't mean my references are going to get messed with that
shit took ages. And they're done. So yeah, I definitely definitely remember the the pain of
referencing the Harvard referencing. It's not it is less bad if you use latex. But
(34:29):
but I didn't for my first draft, I just fucking bashed it out in one caffeinated psychopathic
effort, 20,000 words or whatever in Google Docs and then had to refine it down to 10.
And then reformatting latex, which was stupid. Just right. If okay, so here's right pro tip.
(34:49):
Anybody listening to the podcast, you're going to do any kind of academic work, even undergraduate,
to be perfectly honest, just just start by learning latex and and and just do it in latex
editor. Just don't don't bother wasting your time because you can just get preformatted
(35:10):
Bibtex files of every paper that's published in any kind of non Mickey Mouse journal or even ones
that are independently uploaded to archive. You can just get a Bibtex reference and stick it in
and move on with your life and overleaf is a good way to start. It's free. It's online. You can
access it via your institution. Don't be like this guy and write the whole fucking paper and then
(35:35):
have to refine it and then have to reformat it. It is demoralizing as fuck. Don't do that.
Little D says just use chat GBT to convert to latex.
That's a horrifying suggestion. Please don't do that.
Yes. So on on the subject of chat, GBT, recent subscriber right here, two thumbs up
(36:03):
and have also started using co pilot. And I don't know why I didn't jump on the train earlier.
It's quite handy that shit. It's glorious. It makes everything well, not everything easier.
I think that the real foot gun at this point is having the knowledge to know when things are wrong.
Yeah, that is definitely, definitely a point because it is quite frequently will feed you
(36:26):
some bullshit. But yeah, you know, very confident. I mean, that's the point, right?
Yeah, but it's good for scaffolding. So and it's good for small changes. So, you know, say,
for example, ginger to is quite confusing when you have a large ginger to template.
(36:48):
And say you want to like change or add in like a say you have a really long template and it's
it's got a bunch of loops in there to like bring things in depending on if they're defined and that
type of stuff. And you want to add in like another check, you could just select it all and tell it
to add in another check for this thing. And if you're like pretty precise with your prompt,
(37:13):
it'll do it and do it well. If you're not precise with your prompt, it may do something else.
I mean, it sounds like you just created something very bad, which maybe is the whole point of why
it's fine because it's fine. What do you mean created something bad? Explain yourself.
So you just described nested loops, I think, which, you know, generally pretty well,
(37:39):
ginger is a templating language is not aware of where ginger is now. I used to work in Python.
Yeah, but, you know, is there a different way to do it?
Well, you can obviously abstract out some logic so that you don't have to do nested loops and you
can also do things like list of comprehensions and intermediate functions. Like there are lots of
(38:01):
other ways of breaking out abstractions of things that need computing or checking in kind of because
a lot of what you would be doing is guard clauses. And actually, it sounds like you're
doing unnecessary iteration. And that sounds like exactly the sort of thing that you can probably
get chat GPD to come up with a solution for you. But it's going, it's back to Schultz's foot gun
(38:24):
thing. In fact, it's exactly I am just reiterating at longer length and more confused because I have
worked out a 10 hour day and then gone to a networking thing and then come on the podcast.
And I have no words left. I am just copying Schultz's homework of what he already said,
literally a minute ago. So I'll just say Schultz already said the thing about the foot gun and
(38:46):
the wrong and the okay. So, uh, Cope, oh, it's great for doing dumb shit then if that's what
you're into. I think it's just dependent on the environment you're in, I guess. Well, when you're
templating in Ansible, you only have so many fucking options. And you know, these, I don't see a way
around it. I'd love to have you take a look at it and you can tell me. I think Schultz had it.
(39:09):
Schultz had a point just there before I interrupted him. Well, I was going to go off on a bit of a
segue. So the nulls continue on this topic. I'm totally not doing it. Segue way out. Well, this
is irrelevant to the podcast, but I'm still kind of connected to my, um, my alma mater and chat GPT.
I think it's really damning for oncoming software engineers. Um, they won just the existence of
(39:36):
it means that junior engineers kind of having a hard time getting jobs, but two, they don't know
enough to know not to trust chat GPT and they can't kind of like intuit what's wrong with their answers,
like chat GPT's answers. Um, and I think that that's a serious, um, serious problem that
(39:56):
incoming software engineers have to deal with. Yeah. Fuck it. That's okay. Yeah. That's, I was
literally having a variant of that conversation earlier with, um, because not with the university
links, but I still do consulting and things in like trad tech stuff. And as somebody who's done
a lot of hiring before, I guess, like, especially like former colleagues and whatnot, they've just
(40:19):
gone on the traditional career and engineering career ladder while, you know, we were off being
DGNs over here. Um, but I guess I still, I still just about can, can like resuscitate
somebody else's opinion as my own to, to nod along with these conversations of people who have
responsible real jobs in tech. Um, but that is like something that has come up a fair bit is that
(40:40):
it's that, um, yeah, like that intuition, like you said, that you kind of, you just build up as
like work it like, um, uh, what's the term for journeymen, you know, like when you had craft
people in like olden times, you had like a novice apprentice and then you had a journeyman and
they would literally journey between different master crafts people and like learn from them
(41:01):
and build up their own intuition of how to create the outcome they were looking for, whether that
was a chair or a desk or a horse drawn carriage or whatever. And it was only when they became a, and,
and, and, you know, the, the idea of a master work, you know, it's sort of like, you cannot be
a master crafts person until you come up with your masterwork, right? And it's not possible
(41:24):
unless you go through all the pain of the iterative, iterative steps to build up the,
the reasoning of how to practically do these things. Because there's obviously book learning and,
and yada, yada, yada, but a lot of it is intuitive, especially in the past. And a lot of people have
made the connection between software engineering or programming, you know, programming as it is
outside of a big company, I suppose. And that kind of style of, of learning and progression. And
(41:49):
certainly at least for me, that's the, the progression that I went on from junior to senior
to whatever took that shape. And I saw it in people that I managed and mentored and whatever.
And I think it's quite a good model. And, and, and actually, yeah, and that's the thing I kind of
worry about, like, I suppose, with those sort of things, it's like similar to what you were
(42:10):
saying, Shorsi, is that like, if you, if you don't build that intuition early, then you,
you really don't have a foundation for much to build upon because you're always just building
feedback loops for intuitive processes that are usually below your conscious brain. And then the
bit that the conscious brain is doing is usually the feedback loops around, about learning and
(42:33):
problem solving, which are like how you, how you gather enough information for your brain to then
synthesize the solution. And that's quite a different problem space from the one that your
subconscious is working on. And the subconscious one only gets fed by experience. And chat, chat
gpt is basically short cutting that intuitive leap, it's short cutting that aha moment.
(42:58):
Yeah, exactly. I think that coding book came through an interesting example of that where I
know there's a lot of criticism coding boot camps, but some of the best engineers I've encountered
came from them. And I think that a coding boot camp will be a particularly bad spot for AI usage,
(43:20):
not to their fault, just that I've gone through them as like, kind of as an exercise to see what
it looks like. And a lot of them, you know, they do kind of advanced topics fairly quickly, and
they don't delve deep enough. And so it would be such an easy out to just be like, okay, well,
I'm not seeing this, but how do I, I don't know, convert HTML using CSS to make this widget bigger
(43:47):
as an example. Yeah, so like instead of instead of thinking harder on the on the thing, you just
get something else to do it. Well, but also sometimes it might genuinely be beyond you at that
point. So around the time I got my first job as a programmer, I mainly worked on front end stuff,
and I did some Ruby on Rails stuff. But you know, Ruby on Rails is relatively straightforward
(44:10):
within its own framework, like you can do a lot of back end stuff without having to know
how it all fits together. And I kind of knew I had this like big gap in my knowledge, right? I had
this very applied knowledge of like, decent HTML CSS JavaScript, okay, Ruby, okay. And then
I did one of those, I did one of those, you know, CS 101. There wasn't the MIT course where it might
(44:32):
have been like Udacity back in the day. I can't remember. And I remember doing like the basically
you work through an assignment, the whole course was gradually worked through an assignment,
where you essentially built a first a recreation of the original Google page rank algorithm.
That was what I was working up to. And it was done in Python. Come full circle to what would I
(44:57):
know about Python anyway. And I really, really distinctly remember like the thing that sticks
with me about that process is all the failures, right? Because and what was really interesting
about it is like you were saying shortly about when you get dumped into a topic area where it's
it's actually the gap between where you are and where the topic area is is potentially too wide.
I was working as a developer at that point. And I couldn't, there were some bits of that of that
(45:23):
algorithm, especially more complicated parts, which I actually couldn't couldn't figure out.
And if there had been chat GPT that I probably would have been able to get from a to b and solve it.
But I wouldn't have had any understanding. And what ended up happening was some of the final I
think I did get through it. And then I went on to the follow up and I literally got brickwalled by it.
(45:45):
And I couldn't do the next iteration of it. So I think I finished the CS 101, but I didn't
finish the 252, which was the follow up. And then many several years pass. And then I'm working on
a problem, which is a graph problem when I work for a data consultancy. And I realized I was working
on the problem that I had seen and failed at and never finished like several years before.
(46:08):
And obviously, in that context, I actually kind of by that point did know what I was doing,
solved it because it's part of my working day had to be done. And then just went back and looked it
up out of curiosity and was like, fuck me, no wonder I couldn't figure this out at the time.
But then had, you know, a very belated aha moment looking at it and going, oh, yeah, no,
actually, no, but I see, I see where my reasoning was going with what I've tried to do here. And
(46:32):
it was I was going completely the wrong direction. That's not how you solve this. Fine. But it's that,
you know, sometimes, sometimes learning something can be in the post for a very long time. And
maybe this is just me. I'm a very slow learner. So this is like very difficult of like, I don't
understand something. And then like years later, like fucking cosmos pre commits, man, that just
(46:57):
for the longest time, like, well, I was actually working on Juno Core just and then like years
later, you just like, oh, yeah, okay, mem pools. Everybody keeps talking about mem pools. And it's
not when you it's not until a long time later when you're like, but what are the essential
components of building a blockchain? And you're like, you build a blockchain in your head from
first principles, and you always end up with a mem pool. And you're like, you know, there's a lot
(47:19):
of things in life like that is what I'm saying. And if you can shortcut the, the reasoning required
to understand why you need a mem pool, you end up building a blockchain without a mem pool and
just be like, oh, fuck, we need a mem pool. Like, I don't know, that's probably a bad example.
Well, I get it though. It's like, you need to have that difficulty in order to overcome it.
(47:41):
And not being able to let things simmer because you get an instant
gratification answer is not doing yourself like you're doing yourself a disservice,
to a certain extent, like at this point, like, no, I'm not criticizing you directly in any way,
shape or form. I'm not saying that you shouldn't be using chat tbt.
This is like 25 minutes. This has been purely criticizing enough the last 25 minutes.
(48:02):
Yeah, exactly. All for all for using chat for co-pilot and chat tbt.
Now, for us, for us running a business and we need to use as like a periphery of what we're
actually doing, like, who cares? Right? Like, I'm like, we it's solving something.
The work that works. Exactly. Yeah. But for someone that's still building up their core
(48:24):
strengths in order to create their career, that's where I think it's it can be a damning tool.
Yeah, I mean, I'm not trying to make a career in fucking programming, but
I mean, I make a career in degeneracy, baby. Yeah, it suits our purpose. Even if it's doing
the wrong thing, even if I'm doing the wrong thing, it works perfectly for what I need it for.
(48:44):
So, you know, I don't have the benefit of being a programmer for the last fucking 20 years. So,
if I do things a little wrong, fuck it. I like it because it works.
But I use it as well. I have subscribed to it for like the last, I don't know, 12 months, 18
months. I use it mostly for like APIs. I don't like researching API. I'll just be like, Hey,
(49:07):
how do I send this over here? And then it'll be like, boom, I'll be like, cool.
I find it most useful for other people's API for shit documentation. God damn.
I find it most useful for like brainstorming ideas on how to solve a problem anyway. I use it a bit
in VS code for helping me just quickly change shit. But in like, you know, if I got like a more
(49:33):
complex problem with what I'm trying to do, I'll just talk to it for like an hour and say, you
know, start off with my issue and then explore with it the options it thinks is the way to do
things and then maybe come up with a scaffold of how I want to attack that problem and then start
trying to attack it. But I find it quite handy for that type of stuff. Not so much co-pilot,
(49:56):
but definitely chat GPT. Yeah. So I mean,
So you basically pair programming with it?
Yeah, because, you know, I'm not a, I don't have the benefit of years of programming and I'm
just building. So it's quite handy for that. Thinking out a problem, I guess, thinking.
I've got a, out of pure curiosity, this will take longer, but put a little,
(50:23):
put a little mascot or a duck or something on your desk.
And ask the duck. I knew that's where that was coming. I was going to say,
this sounds just like rubber ducky debugging the game.
It would take longer, but see whether you get to the same solution
and how that, whether there's a difference in satisfaction.
I really can't talk today. Satisfaction. I'd be really, really interested because like,
(50:45):
I've got literally got on my desk here, I talked to Totoro and in Totoro's little bag,
I also have the Buddha in case Totoro and cutting it. And you need to,
you need to appeal to a higher, a higher authority. You ask him, you ask the Buddha,
and if you need the highest authority of all, you got president business,
(51:06):
you got president business in case you really need with a little light.
This is what my Tanooki's for.
Exactly. You're just like, president business, could you help me? I can't solve this thing.
Okay. So what is asking the thing going to get for me? Explain.
Oh, right. So the point is you're always,
the solution is always going to come from your hind brain, right? It's going to come from an
(51:29):
intuitive leap. So what you're always trying to do is feed data to your brain to synthesize the
problem and then make the intuitive leap to solve it. So by ideating out loud, you iterate towards
a solution even in the absence of somebody talking back to you. So my thesis here, right,
is that chat GPT is useful, but the majority of the usefulness is your own brain chewing the problem.
(51:55):
And I'm just, I'm just naively curious, I suppose, whether you could get the same result
from describing your problem out loud to an inanimate object and letting your subconscious
figure it out. Because that's what junior developers were taught to do when, when I think
in some instances, yes, and other instances, no, because I don't necessarily have all of the inputs
(52:19):
required to be able to make the intuitive leap. For example, like there might be a technology
or a software that I'm just not aware of and wouldn't be able to arrive at on my own, a method
even that's employed by a particular software or package or something that I don't know about.
(52:40):
I might need to do something, but not think that it's possible with what is already sloshing around
just because I don't have the breadth of knowledge already to arrive at that. But if it's just a logic
problem or something like that within a program, probably, yeah, I would assume that type of stuff
(53:00):
I'd probably just think on anyway, while I'm like taking a shit. So I'd ask the wall, I'd ask the
Dunny.
Always ask the Dunny if you're stuck.
Yes, shits quite often take an hour and a half instead of 10 minutes just because I'm staring
at the corner of the shitter thinking.
And we too could have this problem if you decide to take the exciting career choice of being a
(53:24):
validator.
Popular Podcasts
New Heights with Jason & Travis Kelce
Football’s funniest family duo — Jason Kelce of the Philadelphia Eagles and Travis Kelce of the Kansas City Chiefs — team up to provide next-level access to life in the league as it unfolds. The two brothers and Super Bowl champions drop weekly insights about the weekly slate of games and share their INSIDE perspectives on trending NFL news and sports headlines. They also endlessly rag on each other as brothers do, chat the latest in pop culture and welcome some very popular and well-known friends to chat with them. Check out new episodes every Wednesday. Follow New Heights on the Wondery App, YouTube or wherever you get your podcasts. You can listen to new episodes early and ad-free, and get exclusive content on Wondery+. Join Wondery+ in the Wondery App, Apple Podcasts or Spotify. And join our new membership for a unique fan experience by going to the New Heights YouTube channel now!
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
On Purpose with Jay Shetty
I’m Jay Shetty host of On Purpose the worlds #1 Mental Health podcast and I’m so grateful you found us. I started this podcast 5 years ago to invite you into conversations and workshops that are designed to help make you happier, healthier and more healed. I believe that when you (yes you) feel seen, heard and understood you’re able to deal with relationship struggles, work challenges and life’s ups and downs with more ease and grace. I interview experts, celebrities, thought leaders and athletes so that we can grow our mindset, build better habits and uncover a side of them we’ve never seen before. New episodes every Monday and Friday. Your support means the world to me and I don’t take it for granted — click the follow button and leave a review to help us spread the love with On Purpose. I can’t wait for you to listen to your first or 500th episode!