May 15, 2026 • 46 mins
This week has it all: geopolitical FOMO, major AI deals, more courtroom drama, and a hacker group that just won't quit. Reed Albergotti (Semafor) breaks down why AI tokens are the new oil, and why Anthropic is buying compute straight from SpaceX. Nitasha Tiku (The Washington Post) takes us inside the Elon Musk-OpenAI trial, where Shivon Zilis, aka the “Elon Whisperer” and the mother of four of Musk's children, finally took the stand. And Pulitzer Prize-winning journalist Dexter Thomas gives insight into how the hacking of education tech platform Canvas could still affect college students, even after Canvas’s parent company says a deal was reached to delete the stolen data.
Additional Reading:
- Nvidia CEO joins Trump in China despite ‘awkward’ politics | Semafor
- Anthropic-SpaceX compute deal shows how tokens are taking over the economy | Semafor
- Shivon Zilis was Elon Musk’s ‘bridge’ to OpenAI. Now she’s entangled in his lawsuit. | The Washington Post
- Instructure Pays Ransom to Canvas Hackers | Inside Higher Education
Download SAILY in your app store and use our code techstuff at checkout to get an exclusive 15% off your first purchase! For further details go tohttps://saily.com/techstuff
See omnystudio.com/listener for privacy information.
Listen
Watch
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:11):
I wanted to start this week by talking about fomo
and asking each of you about any recent experiences of fomo.
And I have some fomo. In fact, it's existential fomo
because read your longevity conference today.
Speaker 2 (00:24):
Yes, it's called the Vitalist Bay Conference. It's great.
Speaker 3 (00:27):
It's gonna shame me into being more longevity ish. I
don't know, Tank Sure, have.
Speaker 1 (00:33):
You any bad fomo recently?
Speaker 4 (00:35):
Oh?
Speaker 5 (00:35):
Man, you know I had some this morning? Actually. So
there is a graffiti artist who has been just absolutely
tearing up downtown LA and I missed seeing one of
his late his I have an idea of who it is,
but I miss seeing one of his latest works by
(00:56):
probably a few hours. It was being painted over. As
somebody told me about it. I went to the location
and somebody's painting over it. So I missed it.
Speaker 1 (01:04):
I missed it, Natasha, anything from you?
Speaker 6 (01:07):
Well, this week was maybe the first time I ever thought, oh,
what if I got to be a White House reporter
and like be in China with Jensen Huang and and
uh Trump and see how those negotiations went, be the
only woman in the room. So yeah, otherwise it doesn't
really appeal to me.
Speaker 1 (01:27):
Well, you said, that the two magic words for this
week's episode intro, which is Jensen Huang and fomo, because
he almost didn't make it, right.
Speaker 3 (01:35):
Rud, That's right. He wasn't going to go. It was
too politically touchy.
Speaker 1 (01:41):
Now it wasn't He wasn't He initially not invited and
then he literally found a way to get to Alaska
and jump on F Force one mid trip I mean
not one plane to plane, not like top gun.
Speaker 3 (01:49):
But yeah, no, it wasn't a mid air transition, although
I feel like with that leather jacket he probably could
do that. But you know, no, it was like if
the President said, hey, I want you to come to
China with me, hop on Air Force one. We're going
to stop in Alaska, and you're a billionaire, you just
you just take your plan to Alaska.
Speaker 2 (02:08):
I guess hop on No big deal.
Speaker 1 (02:11):
Why wasn't he because so Tim Cook and Elona there.
Why was Jensen not originally included in the in the past, Well, he's.
Speaker 3 (02:18):
Been advocating to you know, to sell basically to have
like basically no restrictions on on chip sales to China,
and he's got some of what he wanted, and there's
and there's a conflict there with the China Hawks who
think that's a bad idea. It's also there's just a
ton of scrutiny around him and how close he is
to the president. Meanwhile, AI, you know, AI is becoming
(02:41):
very unpopular politically, and it's it's sort of becoming a liability.
So it was like, Ah, this is going to cause
some awkward conversations. Let's not bring him. And then I
think that became a bit of a news cycle. Maybe
it was on TV and the President saw it and
was like, actually, no, I want that guy. I want
leather jacket guy. Bring him.
Speaker 1 (03:02):
Let's get into it. Welcome to Tech Stuff. I was
Volosha and this is the Week in Tech where I'm
joined by three of the most plugged in tech reporters
to break down what's really happening right now. Today we're
joined by Reid Albergotti from Semophore pul it, a prize
winning journalist, Dexter Thomas, and Natasha Tiku, tech reporter for
the Washington Post. Read I want to start with you.
You told us on this show back in March about
(03:24):
this brewing battle between the AI companies who has most
compute and then recently wrote about how Elon Musk's space
Xai is licensing its computing power to Anthropic. So I
want to talk about that deal. But before we get there,
you've been you've been token maxing on this show for
(03:44):
a while, and I'm just wondering if you could give us,
give us your explanation of why why we should all
be paying as much attention to the new token economy
as you are.
Speaker 2 (03:54):
Yeah, I mean, yeah, I don't.
Speaker 3 (03:55):
I don't think i'm token maxing compared to you know,
a lot of people out here in the day area.
But I do use these things every day, and I'm
finding I'm getting a lot of use out of them.
And I'm also sort of aware, just based on my
own informal polling that you know, I'm actually a pretty
early on the adoption curve here. Like I don't, I
don't think we've really seen anything yet in terms of
(04:15):
how many tokens people are going to be using when
when this stuff really catches on and gets better and
easier to use. And you know that just to me
says a lot about this token economy, like the demand
is going to be there for a long time.
Speaker 1 (04:29):
But I want to ask you about that specifically because
you've made this argument quite you know, consistently for the
last few weeks. I haven't really seen anywhere else and
we haven't actually paused to unpack about tokens as a
kind of new currency or commodity, Like are they like bitcoin?
Are they like US dollars? They like gold? Are they
like cloth? Are they like you know, oil? Like what
is the what is the most familiar form to understand tokens?
Speaker 3 (04:51):
Yeah, I mean they're kind of like bitcoin, but they
actually have utility, right like people people need them and
want them, and you sort of realize that when you
try to use you know, the endthrop products, because you
run out of tokens pretty fast, and then you find yourself,
you know at qwo am sort of like a heroin addict,
like just wanting more tokens, you know, And I think
I think that's like to be really indicative. And and
(05:13):
this elon the whole XAI or SpaceX, whatever you want
to call it, deal with Anthropic where you know, XAI
found itself actually with some extra tokens, and instead of
that being a problem, they just unloaded those two Anthropic
which was short of tokens, and that to me was like,
oh wow, like these things are if their general purpose
they're transferable, they are. They are really becoming like a
(05:37):
like a commodity that you can just easily crede.
Speaker 2 (05:40):
Just fascinating.
Speaker 1 (05:41):
I want to I want to come to Dexter, who
is chuckling about about you shooting up tokens late at night?
Just just before that, really I had one question for you,
which is which has been burning on my mind actually,
which is how much surplus value is there in token creation?
I mean, in other words, like today, when you assess
the price of building a data center, the extric see
that goes into the data center, can minting tokens be
(06:03):
a profitable endeavor or the tokens cost more to make
than their worth today?
Speaker 3 (06:08):
Well, I think it's hard to I mean, these companies
are losing money, so you might say, oh, well they're
actually you know, they're actually not. It's not a profitable business.
But you know, I don't, I don't exactly. I mean
there's a lot of I don't exactly know where the
all the costs are in that process. But we do
know that with compute, it's just it always the price
(06:28):
always comes down. That's just how it works. And you know,
I don't. I don't think we have we want to
spend the whole show like talking about about going into
like the weeds of that. But I think it will.
It will become a very profitable business. But there's also
like a curve where while the demand is like extremely high,
I think it's going to be very very profitable, and
(06:51):
Xai can say, oh, yeah, like this is this is great,
like we'll just unload our our tokens and it works
for everyone. They can do that now, but eventually this
is going to become a mature business, and just creating
the tokens will be a profitable, but like a much
smaller margin business than if you're offering a service on
(07:11):
top of that. So nobody wants to be just a
dumb token provider. That's not what I'm saying at all.
But but I think for right now, the demand is
so high that it actually is a pretty pretty good business.
Speaker 1 (07:21):
To text to you chuckling about about reads a description.
What's what's been your interaction with token mixing to date?
Speaker 5 (07:27):
Oh my god? Well, I mean have we all heard
about the Caveman token hack or the Chinese token hack? Okay,
all right, I'm very sad to introduce this to the group,
But basically, this is a It's really interesting that most
in general, most people don't necessarily know how the inner
workings of AI works. But we're starting to see just
(07:50):
end users understand this to a degree, right right, And
people have come up with these kind of like chat
gbt slash claude life hacks. People are speaking to chatbots
and having the chatbots speak to them as briefly as
possible and just telling it be as terse as possible.
(08:11):
And the absolute most terse you can be is what
people have called like Caveman mode, which is don't use
any don't use prepositions, just say job done, need fix
like things like that, and you speak to the chatbot
like that it speaks to you. But it also just
so happens that English is kind of an inefficient language
(08:31):
because we have a whole bunch of letters that don't
actually do much. Chinese is much more efficient in that
it's able to pack a lot more meaning in, for example,
just one character, and so Chinese users have realized, wait
a second, I can just why would I talk to
this thing in English? I can just talk to it
in Chinese. The problem is It doesn't necessarily always work
as well as you might hope, but you know, people
(08:53):
are hoping to eke out an extra ten, maybe fifteen
eighteen percent in terms of of token usage by either
speaking like a caveman uh and saying good job, bad job,
please fix, or speaking a literal other language so they
can not be sitting there at two am waiting for
their Claud tokens of reload. Maybe they're sitting there at
(09:15):
three am waiting for the Claud tokens a reload.
Speaker 1 (09:18):
But there's an interesting there's interesting cultural tension here because
on the one hand, if you're paying for your tokens,
you want to be in caveman mode and not burned
through too many. But if you work for a tech
company like I guess that you want to yeah, but
precisely building the most inefficient possible ways of working to
get through as many tokens as possible yea, which is
just wild to think about.
Speaker 2 (09:40):
In my experience.
Speaker 3 (09:41):
That's not actually that doesn't actually end up saving you
that much that many tokens, because it's like you're then
just having to do more, you know, more prompts to
get to the you know, you want to be like
as detailed and as specific as possible when you're starting
out in a project. But it is an interesting one.
And like, you know, I had an experience. I built
a website the other day and I was sending it
(10:02):
around to friends and each time they use the website,
it was costing me money, right, And so I had
it build the dashboard to see how much money this
website was costing me, and it was like about seventeen
cents every time someone was using it.
Speaker 2 (10:16):
And so then I had it.
Speaker 3 (10:17):
I was like, just run an evaluation and you know,
find a better, more efficient way to do this. Here's
a bunch of different APIs, try them all out, and
it brought the seventeen cents down to like a fraction
of a penny and it still works.
Speaker 2 (10:31):
And I was like, oh, that's pretty cool. You know,
there's very.
Speaker 1 (10:33):
Interesting I got. I got LinkedIn shamed. I wish there
was a regular experience by somebody telling me that as
a founder, if I wasn't spending as much a month
on tokens as I am on rent, I'm already doing
myself a huge disservice. So goals.
Speaker 5 (10:48):
This is the really interesting thing, though, is I mean
you brought up bitcoin, which obviously this is not a
one to one comparison. But the really interesting thing is
there are people who were really really into bitcoin and
they don't think or ethereum, but whatever cryptocurrency you want
to talk about. They don't think of their net worth
in terms of dollars. They think of their net worth
(11:09):
in terms of bitcoin. How much bitcoin.
Speaker 1 (11:11):
Doner, the conversion is not they're not. They're not. They're
not even monitoring the exchange rate.
Speaker 5 (11:14):
Or they'll tell you that they're not right. They are.
They know exactly, they know that they know the Bitcoin
to US D ratio, they know what it is. But
they'll tell you. Don't think of your money in terms
of fiat. That money is dirty, it's the government money.
Think of it in terms of bitcoin. And it's very
interesting that you're having people think of tokens as a
(11:35):
unit of value. That that to means fascinating.
Speaker 3 (11:38):
I just think of how much I owe my mortgage
that's being in California. But go ahead, I didn't mean.
Speaker 1 (11:45):
Well, Nontas, I just wanted to say to you. I mean, obviously,
the prompt for this conversation, so to speak, was this
was this space Xai anthropic deal which kind of interesting
right space XAI. Elon was praised for building these massive
data centers early, and everyone was jealous and apparently like
other companies were flying drones over his building sites because
he was like in maniac mode and doing construction much
(12:07):
fast than everybody else. But then, of course his product
isn't like as useful in terms of actual AI than
Anthropic or Open ais Le Google, So it's kind of
interesting irony there. But also I think three months ago
Elon said that Anthropic was misanthropic and it was evil,
and there's a break clause in this contract if he
decides to Anthropic it evil again. So I mean, do
(12:29):
you see another lawsuit brewing here? Like what hell is
this going to go down?
Speaker 6 (12:34):
I I mean it just it just goes to show
you like Compute brings all men together, you know, they
like when it's when it's a deal of convenience. Elon
conveniently changed his mind about who's okay being in charge
of AGI. I just find his like word play to
(12:55):
be so bizarre, like why fixate on the name of
open AI being open or misanthropic, Like it's just.
Speaker 1 (13:07):
It's quite trumpy and right, It's like a very very yeah.
But I mean it's actually been reporting on Silicon Valley
for a while. I remember when like when you know, Google, Facebook,
you know, et cetera, were and Apple were kind of
all becoming behemoths in like twenty ten era. Eric Schmidt
(13:28):
was on Apple's board i think, and there was quite
a lot of like sort of circular. It's a small community,
and it felt like a rising tide was, you know,
raising all boats, and then all hell broke loose here.
It seems like it's not sequential like arising. Everyone feels
like a rising tide is raising all boats and they'll
work together, but all hell but it's also breaking loose simultaneously,
which is a kind of an interesting conundrum.
Speaker 6 (13:49):
Yeah, it's a tangle of ties, but tenuous ties that
could you know. Also, you know, we're all we're all
living in a billionaire's world where they're ego you know,
even sometimes more so than their commercial needs, are dictating
our government policy, our geopolitical rivalries. You know, all of
(14:15):
a sudden, Trump might come back, like depending on what
happens with Jensen Wang and you know in China, might
come back saying like, oh, you know, she's our buddy, Now,
everything's fine. He was already moving away from you know,
talk at least about selling weapons to Taiwan and protecting Taiwan.
(14:35):
So truly like five to ten men control the world.
Speaker 1 (14:41):
Yeah, I mean we're going we're going to talk about
Chavon Jis after the break. But it was interesting. He
brought three hardware people, right, I mean, Tim Cook, crom Apple,
Jenson Huang from a video you know, must from TESTA
and SpaceX and others. But but none of them. But
neither Samilman nor Dariama Day nor you know, the kind
of fundamental AI model we're not included in the in the.
Speaker 6 (15:02):
Picnic, Yeah, he had some you know, like more mature,
grown up, regular business folks like that. I'm sure he
sees more often in New York. But I don't think
you're going to bring Darya Amiday, Considering you know, he
is a huge China hawk and his rhetoric around their
(15:24):
authoritarian government. I think he would definitely be off the list.
Speaker 1 (15:28):
We're going to take a short break now. When we
come back, Atash will give us the latest updates from
the Elon Open Ai trial and read you're off to
shoot some peptides.
Speaker 2 (15:37):
Yes, I am good luck.
Speaker 1 (15:45):
When you're traveling abroad, wi Fi is always an issue.
Maybe you can't find it, or maybe you can, but
the Wi Fi network that's available looks a little bit
sketchy and makes you think it might not be safe
to connect to. I've certainly had that feeling, and it's
why having a local simcard on your phone can make
all the difference. But you don't have to literally buy one.
(16:06):
That's where Sale comes in, an e SIM service that
is as simple as downloading an app, and that gives
you instant Internet access wherever you're traveling. That is safe
and secure Internet with built in cybersecurity features like web
protection and ad blockers, and no need for local Wi Fi.
All you have to do is download the Sale app
on your mobile device and choose the plan that works
(16:26):
best for you. Once you install the e sim, Sale
will be instantly activated, so you can start your trip
off right. Download salely in your app store and use
our code tech Stuff at checkout to get an exclusive
fifteen percent off your first purchase. See details in the
podcast episode description box. Welcome back to tech Stuff, Natasha.
(16:47):
Let's move onto your story. You've been filling us in
on the details of the Elon Musk open Ai trial
remind us what the trial is about and why Shivon
Gillis is such an important character.
Speaker 6 (16:57):
So this trial is about, as Elon must put it,
you know, did open ai steal a charity? Did they
take this nonprofit company which was supposed to be you know,
a counter force against Google or China or you know,
some other large entity having control of ai in the future.
(17:21):
Did the idea of them getting money from Microsoft, you know,
getting billions of dollars from Microsoft? Did that change the dynamics?
Did Sam Altman, you know, enrich himself at the cost
of humanity? Is basically Elon's argument. So what's central to
that are the negotiations that the board had with changing
(17:42):
the structure of the company from a nonprofit to a
for profit. And Chavon Zillis was on the board at
this time. You know, she approved those decisions to get
funding from Microsoft. So she is, in a way like
a key witness for open Ai for the defense.
Speaker 1 (18:03):
Which is interesting. So she's she's the mother of four
of his children. She was on the board of open Ai,
but was also communicating with Elon during that time. But
now she's essentially a witness for open AI. Was she
called by by the defense or how did she end
up on the stand or what was it? Like what
happened there and she.
Speaker 6 (18:22):
Ended up on the stand because so much of her
communication was you know, in the discovery for this trial,
many of her notes, her personal text messages. They're clearly
very close and clearly her loyalties which came out in
her testimony, Clearly her loyalties are towards Elon Musk.
Speaker 1 (18:40):
I want to ask you more about about the trial
and what's come out during the during this her on
the stand, but you know you had this phrase before
the break about what was it tokens bring all men together?
What was it? I mean, this is compute brings all
the men together. I mean this is interesting. She's a
female board member. What are the kind of gender dynamics
of hivon the bros.
Speaker 6 (19:01):
Yeah, I mean you see her, this is like the
one woman right like that is that has become a
central figure in this trial. And what she's doing is
essentially kind of you know, in some cases like massaging egos,
smoothing tensions, you know, trying to translate one person's needs
(19:23):
for another. Like I said, like emotional labor for these men,
and you know, it's just it's kind of it's frustrating
to watch. There's one exchange where sam Altman is asking her,
is she going to attend, you know, one of the
rocket events, and she said, oh, no, no, no, no, I
wouldn't even dare ask. And it just, you know, this
(19:47):
deference that she has towards Musk, you know, we just it.
I guess it makes me wonder, you know, how she
describes some of these things to her friends or to
her family, if she does it all. We also learned
on the stand that her until until Business Insider reported
(20:08):
that Musk was the father of her children, even her
own father did not know that.
Speaker 1 (20:12):
In fact, I've read that her first call when they
called her to fact check the story was to her
father and her second was dissembleman, Yeah, what a life.
Speaker 6 (20:20):
I know. I mean, you know, this is a woman
who has tied her future, her personal life, her fortunes
to a very mercurial man who describes her as as
his chief of staff. And sorry, just to clarify too, Chavon.
At the same time that she was on the Open
AI board, she was advising Musque on neuralink Tesla, you know,
(20:43):
his kind of his portfolio across the board.
Speaker 1 (20:46):
Texter, what's your take on with this?
Speaker 5 (20:48):
I keep thinking of, you know, how so many people
think that their phone is listening to them, and yet
they use their phone anyway. And I think a lot
of people are watching, you know, if they're paying attention
to this at all, the you know, this sort of
courtroom drama thing happening and feeling like, well, they go
at it again. It's as bad as I thought it was,
(21:10):
but here's a few extra details. It's it's sort of strange.
It's strange to watch.
Speaker 6 (21:16):
Yeah, I think if you if you pull back a
little bit too. Part of a large part of what
this trial is is a number of people defining whether
Open AI is adhering to its mission, which was building
AI to benefit all of humanity. And if nothing else,
(21:36):
you get absolute confirmation that it is a handful of
people that are, you know, having their personal perspective on
whether this is good for humanity or not.
Speaker 1 (21:47):
This is the.
Speaker 6 (21:48):
Least democratic process, you know, you have ever seen. There
is one exchange where Shavn is trying to think about
how what are different solutions for getting the money that
need you know, that's not working as a nonprofit. And
she suggests talking to demissabas, the Zee of Google Deep Mind,
(22:08):
and she describes him, you know, like whether or not
he is moral enough, and she says, oh, maybe if
he spends more time around e he'll start thinking about
humanity more. And yet all of the interactions we're seeing
from Elon are concerned with whether or not he's in
control of the company. It's just it, just like brings
(22:30):
the ridiculousness of open AI's stated mission into clear focus.
Whether or not your entire court case revolves around protecting humanity.
These are not people who are talking about humanity. These
are people who are talking about whether or not they
exert control over a product and a technology deck.
Speaker 1 (22:51):
So you put that parenthetical if they're following it at
all in your coined about the time, what do you think?
Speaker 2 (22:56):
Is how nice?
Speaker 1 (22:57):
Hard to know? When you I mean, you know, we
and most of my time reading tech news, right, so
I have been following it, But I don't know. And
this is obviously not the oj trial. But is it
is it?
Speaker 2 (23:07):
Is it?
Speaker 1 (23:07):
Is it something which has permeated the popular consciousness? Do
you think?
Speaker 5 (23:10):
I don't think so, I don't think so, I think,
I think there's a section of us who are definitely
paying attention. But I think, you know, I have conversations
regularly where I'm I'm talking about some of my work
and I say Sam Altman, and I get a who
Who's that? Who's that? And so obviously not everybody's tapped
(23:30):
in with this. I think they. I would hope that
more people would. And I think, Natasha, precisely, what you're
talking about and what's being discussed in this court case
is something that you see and un frankly, a lot
of Sam Altman's projects, I mean, look at one of them,
world Coin, right, which is now Tools for Humanity, which
is you know these orbs right. One of these stated
(23:51):
purposes originally of one of one of his projects was
to provide ubi universal basic income. So AI is going
to take everybody's jobs, and so in order to prevent
the damage or you know, mitigate the damage it will
come from that, let's make sure that everybody in the
world is getting, you know, a basic income. And of
(24:15):
course they started by going to places like Kenya and
scanning everybody's eyeballs. But now if you look at the
language on the website and when they talk about it,
they're not really talking about universal basic income anymore. They're
talking about verifying that you're a human. And if you
keep pressing them about why is that important, eventually you
start hearing about, well, advertisers want to make sure that
(24:38):
the people who are seeing their ads aren't bots, they're
actual humans, even if you know. So, there's a lot
of this shifting, saying one thing and then making sure
that the public thinks, oh, yes, this is what they do,
and then what's actually happening starts to change. And I
would love for more people to you know, be tapped
(24:59):
in aware of this. That is one of the purposes
of the show, of course, but it's tough. It's tough
because it moves so fast. It moves so fast.
Speaker 6 (25:08):
Well, I think I have started to see, at least
with Sam Altman. I think partly because you know, the
nature of how we consume tech information is we treat,
you know, the CEO as like a proxy almost for
the company, and we have this impression that they are
controlling things, often because they do have you know, super shares,
(25:30):
they tip the board in their favor, et cetera. So
I do you know, yes, nobody I think no regular
person is paying attention to the trial, But I think
people do want to get a read on Sam Altman.
I think they do, you know, when the Pentagon stuff started.
They do want to get a read on Dario Amiday.
And with good reason, right like, as you so rightly
(25:52):
pointed out, Dexter, Like what they say is is is
never ends up being like a good indicator of how
their technology is going to impact the public. So you
do need to figure out, you know, how they operate
as a person, you know, what they really care about,
and the language from open AI since the beginning, you know,
they're almost flirting with talking about like wealth redistribution. You know,
(26:18):
they'll like the way that they talk about human flourishing,
or you know they kind of wave at what will
happen when all the jobs are taken away. I think
I think the reason that people are grasping for psychological
reads sometimes on these CEOs is because they know it
will you know, have economic consequences for them in the.
Speaker 5 (26:41):
Future, definitely. And I think the the other thing that
I've been trying to puzzle out, like why is it
that people don't know the names like Amada, Why don't
people know Sam Altman. Some of it is because this
has happened so fast. Like open AI fundamentally for most
people did not exist three years ago. It wasn't a
function anthropic even more so. Right, But I don't think
(27:06):
it's just that. I think, you know, people can name
Tim Cook, people can name Steve Jobs. Of course, people
can you know, go back a little bit further, you know,
Bill Gates, all these sorts of things.
Speaker 1 (27:16):
The thing that Simulman has higher name recognition than Tim
Cook today in America. No, I don't think interesting. I
don't think Natasha.
Speaker 6 (27:24):
I think so really, I all the time, you know,
we have to gauge this because we like fight over
whether or not we can put his name in a
headline for the Post all the time. And I think
he crossed over partly because he does massive media tours,
you know, like well.
Speaker 1 (27:45):
I think made the Sore app which he was the
main star of as well. Right, I mean, you can't
avoid him in April last year.
Speaker 6 (27:51):
And if you you know, like because there are a
number of influencers now who are just sharing clips of
what these guys say. It's just like, you know, nine
hundred days since you know, one of these AI CEOs
doesn't say something insane, so the clips are shared around routinely.
I think it is just a matter of time before
(28:13):
he becomes a Mark Zuckerberg like figure. I mean, maybe
we need a social network movie.
Speaker 1 (28:19):
But I think I think he became co identified with
AI and like his thing is called open AI, not
anthropic or deep mind, right, and the Chatchipet was the
first product. So I would say people don't probably know
dari Amidau's name or Demos Hassebus his name. But I
my uninformed take is the people that Sam Wlman has
become a kind of public figure, and I think, you know,
(28:40):
the fire bombing of his house as well kind of
goes to show the extent to which he's become like
a sort of figurehead or you know, representative of the
whole industry as well as.
Speaker 2 (28:51):
Just a person.
Speaker 1 (28:54):
We're going to get a break now. When we come back,
we're going to hear about a hacker group who stole
data from over nine thousand colleges and universities, affecting two
hundred and seventy five million users. These hackers wanted money
in return for not releasing the data. Stay with us,
(29:18):
Welcome back, Dexster. You brought a story this week about
the recent hacking of an education platform called Canvas. Yeah,
what happened to you break it down from the beginning.
Speaker 5 (29:27):
So Canvas, with those who don't know, is what's called
an LMS, so I think a learning management system. And
so basically it's the way that a student interfaces with
their classroom digitally. So a teacher might you or a
professor might use it to assign homework, to host group discussions,
to give tests, and maybe most importantly, or to post
(29:50):
readings things like that. Maybe most importantly we're going to
talk about this. You can send messages to your professor
via essentially this portal, and so.
Speaker 1 (30:00):
Slack at tech slack kind of kind.
Speaker 5 (30:02):
Of Yeah, it's a bit like that for those of
you Slack. And so this is if you know somebody
who is going to university or college in North America anywhere,
or you know somebody who has been to college in
the last decade, has some change, there's a very high
chance that they've used this. And I say that because
(30:23):
Canvas holds about forty percent of the market in North America.
Speaker 1 (30:29):
So this is a that's a huge two hundred and
seventy five million number.
Speaker 5 (30:32):
Comes from precisely. So let's talk about the hack. On
April twenty ninth, Canvas was first hacked, and so the
basically the company that runs Canvas is called Instructure. So
April twenty ninth, this group called Shiny Hunters, which is
a Pokemon reference, but they're also very well known, pretty
(30:52):
notorious for hitting a whole bunch of other companies very
successfully and very lucratively. I might add to the tunes
of millions of dollars they hit group. They're all over
the place. They're all over the place. So, and this
is the thing about hacker groups. I mean, if we
knew where they were, it'd be much easier. They probably
wouldn't be operating anymore. But so April twenty ninth they
(31:17):
get hacked. May seventh they get hacked again with the
same method. So, without getting too much into the weeds here,
this is a little bit like breaking into the front
door and then a little over a week later breaking
into the side gate. Like they basically did the same thing,
(31:37):
but the difference is on May seventh, basically they posted
a ransom note on the Canvas portal for all of
these students, So students who were trying to log in
to turn in assignments, and mind you, this is final
season at Harvard, at Penn At a lot of really
(31:59):
prominent elite universities get on and they see a ransom
note saying that hey, we have taken down canvas, We've
gotten all the messages, we've gotten all this other information,
and Instructure won't even talk to us.
Speaker 1 (32:15):
So the first breach, the front door, as it were,
was to basically extract all of the data.
Speaker 5 (32:20):
It looks like they got some, and they may have
gotten more.
Speaker 1 (32:24):
It's a little hard hard to tell, but enough to
be able to do the second breach, to say, if
you don't pay us, we're going to release all this data.
Speaker 5 (32:31):
Is interesting, precisely, precisely. Yeah, And so that's when it's
public in the sense of instead of just talking to
Instructure and saying, hey, give us money, now, they're going
directly to the students, and that guarantees that it gets
media attention.
Speaker 1 (32:50):
And they paid up right, I mean, the ransom note worked.
Speaker 5 (32:52):
Yeah. So okay, the answer to the short answer to
your question is yes, Instructure has posted a message on
the website saying that they have reached an agreement. They
didn't explicitly say that they paid the ransom, but we
know what that means. And they also have said that
the stolen data was given back and that they have
(33:15):
received digital confirmation of data destruction, and they say in
parentheses shred logs. I see your face. I'd like you
to say what you're thinking.
Speaker 1 (33:25):
I mean that sounds like BS to me, right. I
mean the statement they gave, by the way, was, while
there is never complete certainty when dealing with cyber criminals,
we believe it was important to take every step within
our control to give a customers additional peace of mind
to the extent possible. I mean, that is one of
the most overqualified, absurd corporate statements I've had in my life.
Speaker 5 (33:48):
There's a lot of caveats in there. Yeah.
Speaker 1 (33:50):
So, but maybe I'm wrong, and maybe I'm being too cynical.
You think maybe they did credibly destroy the data.
Speaker 5 (33:56):
Data is not like your mother's fine china. It's not
something that you pick up and then you take out
of the house and then you give back. It's Anybody
who's ever used a computer before knows that you can
just make a copy of things. That's what computers are for.
So for the tree, yeah, shred logs. A shred log
is just it's basically a text file saying we deleted this,
(34:18):
but if you made copies of it before. There's no
third party audit, there's no forensic trail. There is a
trust me bro from a group of criminals. This is
how they operate. What we have received is not evidence
that all data has been quote unquote given back or deleted.
(34:40):
What we've received is a pinky promise.
Speaker 1 (34:43):
Natasha, I want to ask you about this because I mean,
this is a story which is affected. You know, a
pinky promise is now guaranteeing the security of the data
of two hundred and seventy five million people, basically everyone
in college and everyone who's been at college in the
last ten years, at all of their professors. I mean,
that is functionally like a huge group. And yet I'm
glad you brought this story. Dexter It got I think
(35:04):
four paragraphs in the New York Times and their longer
piece in the Higher Education Journal. But I would say,
broadly speaking, like no pickup. Whereas mythos, the anthropic model
that has the potential to render all cybersecurity mood is
still dominating the headlines. A month on why Natasha, do
you think that the theoretical power of mythos is so
much more interesting to people than the literal extraction and
(35:27):
theft of this huge amount of data.
Speaker 5 (35:30):
I mean, that's.
Speaker 6 (35:32):
That's such a good contrast, right, I think about this
all the time with with like what counts as AI.
You know, there is obviously the media has a bias
towards what is new, and I think there's just a
sense of powerlessness about how our data is used by corporations.
(35:54):
You know, it's it's not sexy, it's not interesting, it's
just you know, everywhere. Then you have to start thinking
about like ad networks and like Dexter is talking about,
you know, how data gets shared everywhere. You know, you
could also be looking if you cared about AI at
like algorithmic bias in banking, you know, in housing, in education,
(36:19):
et cetera. And yet like that's the focus is on
usually a set of five men. Sorry, keep coming back
to the same theme, but that's the way regulators think,
that's the way the media thinks, and so that is
where our attention goes.
Speaker 1 (36:36):
I mean, Dexter, I think part of the big part
of your your body of work is surfacing stories exactly
like this one that are not the kind of a
one New York Times story about who's up who's down
in the power stakes. I mean, what what did this
story say to you?
Speaker 5 (36:50):
Well, I mean, I'll be real with you, this is
actually pretty personal for me because I teach at universities,
and so, just to be clear, the information that they
got was usernames, email addresses, course names, enrollment information, and messages.
(37:11):
That's the most important part right there, the messages. Right So,
Instructure has come out and said, oh, well, you know,
they didn't. None of the course content was compromised. Nobody
cares if homework number seven from Engineering one oh one
gets leaked on the dark web. Nobody cares anyway, Yeah,
nobody cares about that. The crucial thing here is the messages.
(37:35):
Students have extremely vulnerable conversations with their professors, and it
could be anything. It could be my grandmother passed away
and that's why I can't make it into class. It
can be I'm really struggling mentally and I don't know
what to do, and that's why I haven't been in
(37:55):
class three weeks. You know, can we talk? It can
be this really terrible thing happened to me? What do
I do? I've received all those messages, these are numerous times.
These are at every institution that I've taught at, including UCLA,
which is one of the schools that was hit is.
So that's me thinking as a professor here. Now let
(38:19):
me think as a hacker. The hackers have said that
they are going they are not going to resell the data.
There's no agreement to not use the data. I can
tell you exactly what I'm going to do, what I
would do right now. So one of the schools hit
was Harvard. So I'm a member of Shiny Hunters. Let's
say let's say that I'm not going to resell this data.
(38:39):
Shiny Hunters has resold data before. This is something they do.
It's part of their business model. By the way, the
FBI tells ransomware victims to not pay the ransom for
specifically uses like this because A identifies your company as
a mark and B lets people know that this is
a payday. And this is how shiny Hunters operates.
Speaker 2 (39:01):
Right.
Speaker 5 (39:01):
So okay, let me stop thinking as a professor. Let
me start thinking as a hacker. What I would be
doing right now. Let's say that I'm going to abide
by my pinky promise to not resell this stuff to anybody. Okay,
I'll just use it for my personal use. Here's what
I do. Harvard. I know there's a bunch of rich
kids go to Harvard. I have the contents of the messages,
and I have the enrollment information. So let's say I
(39:23):
just grab a list of a thousand students. I know
what classes they're in, and I know which student is
in which class. So I just send off an email
to a thousand Harvard kids and I say, hey, it's
your bio professor. It's Wednesday. You forgot to sign the
student contract at the beginning of the course. I don't
know how I didn't notice this, but if you don't
(39:44):
sign this by seven pm. By the way, it's five pm,
then your grade is going to be registered as an
F and I won't be able to fix it until
the start of summer semester. So you're gonna have F.
You're gonna have an F on your transcript. I know
that you were going for that internship. I know this
because of the messages. So here's here's an email. Just
click this link, log in, fix it, and boom, phishing attempt.
(40:04):
I got your email address, now I log in. Now
I'm in your iCloud. Now I got all your messages,
Now I got all your photos. Imagine what kind of
photos on there. No, this is and I'm not giving
anybody instructions because this is precisely what I would be
doing right now.
Speaker 2 (40:19):
Yeah.
Speaker 5 (40:20):
No, I know. It is one of those things like
I can definitely imagine somebody saying, Dexter, don't don't give
people ideas. No, no, I'm not giving anybody ideas. This
this is what's happening actually right now. Yeah, fire off
a quick message. Out of a thousand kids, maybe three
get caught, pretty good ratio. Fire off a message to
the kid and to the parents, say hey, I'm in
your kid's iicloud two and fifty K. If I don't
(40:41):
have it by Thursday, it goes up to half a mill.
And you could say, oh, well, I don't care. Harvard.
It's a bunch of rich kids, a bunch of you know,
legacy kids. I do know a bunch of poor kids
who go to Harvard. People get scholarships. So somebody is
going to have their life ruined. Multiple somebody are going
to have their life ruined. Remember, they got billions of messages.
They have a lot of information which is going to
(41:03):
be very useful for social engineering, the best kind of hacking.
And and we will never know because unlike Construction, which
is a public and traded company who has to at
least say, hey, something happened to us. We know that
the psychology of this is people are ashamed when they
(41:25):
get scammed, and so they keep it quiet. And so
we will be seeing We will be feeling the ripple
effects of this for years because if I don't, let's say,
I don't get any kids on that first sweep of
a thousand people, try again in three months. The ROI
on this is astronomically great. We will be feeling it.
(41:47):
We'll never see it because it will never get reported.
I've heard I've heard this said. I'm still reporting on this.
But I've heard this called the worst privacy disaster in
the history of educational tech, and I fully agree with that.
There's been I can't think of anything worse. This is
catastrophically bad. If you know anybody who is a student,
(42:10):
or you know anybody who's a professor, at the very
very least, I know you didn't ask this, But I
feel like I got to say this because I just
said a bunch of bad things. I feel like I
got to give you something. If you know anybody who's
a student, tell them change if they happen to use
the same password in their Canvas account as they use
anywhere else, for example, their email, change that right now.
(42:32):
The next thing would be talk to your university and say, hey,
what are y'all doing? Because there's some things universities can do,
and every university is different, every college is different. But
at the very very very least, it's easy to feel
powerless now, and I f really understand that. But at
the very least, if you're sharing passwords across multiple accounts,
definitely go and change that right now.
Speaker 1 (42:54):
Like now, well, that's the way I have time for
this week. We are going to go out on our
final cake a dexter I think you'll us week in
Tech goes to the studenty student and professor in America. Yes,
my worst week in Tech is also higher ed coded,
but quite a lot less serious and somber than yours.
Gloria Caulfield, who was giving a commencement address at the
(43:18):
University of Central Florida, got booed almost off the stage.
Speaker 5 (43:22):
Oh my gosh, yeah reading the room man not a
strong suit apparently of this individual. That was a rough one.
Speaker 2 (43:28):
Let's say we can play the lip.
Speaker 5 (43:30):
The rise of artificial intelligence is the next industrial revolution?
Speaker 2 (43:41):
WHOA?
Speaker 3 (43:45):
What happened?
Speaker 1 (43:48):
Okay, I struck a chord? May I finish? Thanks for
sound that you'd heard this already. But but Natasha, did
you see this one?
Speaker 3 (43:59):
Oh?
Speaker 4 (43:59):
Yes, yeah, I loved how she like when they started booing,
she looked back at the days like, it's just I
don't know. It was just a very like physical encapsulation
of how Voss layer CEO layer management layer works.
Speaker 2 (44:16):
Like the turning around me, like, what's going on here?
Speaker 1 (44:18):
So I know they're in front of you, that's doing
you because of what you said.
Speaker 5 (44:21):
It's it's so yeah, it's so unfortunate. I mean, I
can't confirm. I have no evidence here, but I wouldn't
be surprised if a few tokens were expended writing that speech.
It was a unfortunately sadly out of touch speech and
it went downhill from there.
Speaker 1 (44:39):
Natsher's is chavon you your West Weekend Tech? What do
you say?
Speaker 6 (44:43):
I'm maybe gonna put Musk in my worst week in tech. Actually,
you know, first of all, I think that there, you
know it wasn't covered this way, but the anthropics SpaceX deal.
Giving up this compute that you knew you would need
or you believed you would need because you thought you
are going to be leading the technology, giving that to
(45:05):
a guy you called misanthropic. That's that's definitely down in
the rankings.
Speaker 1 (45:10):
Okay, let's do our best. One sentence only best for
Jensen Hwang for me because he managed to he managed
to make his dreams come true and get on the
trip after all.
Speaker 6 (45:19):
Jensen for sure. And whoever makes Jensen's leather.
Speaker 5 (45:22):
Jackets everybody who is in the audience at that unfortunate
commencement speech because they're they're famous and they can do
something with it. I don't know what do somebody?
Speaker 1 (45:34):
That's it for the we can take. Thank you all
so much for joining us today the Tech Stuff Finals
Veloci and This episode was produced by Eliza Dennis and
Melissa Slaughter. Executive produced by me Julian Nutta and Kate
Osborne for Geidoscope and Katrian Novel for iHeart Podcasts. The
engineer is Kathleen Conti from CDM Studios. Jack Intolely mixed
this episode and Kyle Murdoch wrote our theme song special
(45:54):
thank you to Read Albergotti, Dexa Thomas, and Natasha Tiku
with Practical and Friends of the Paul. And we'll go
for to see you next week
I wanted to start this week by talking about fomo
and asking each of you about any recent experiences of fomo.
And I have some fomo. In fact, it's existential fomo
because read your longevity conference today.
Speaker 2 (00:24):
Yes, it's called the Vitalist Bay Conference. It's great.
Speaker 3 (00:27):
It's gonna shame me into being more longevity ish. I
don't know, Tank Sure, have.
Speaker 1 (00:33):
You any bad fomo recently?
Speaker 4 (00:35):
Oh?
Speaker 5 (00:35):
Man, you know I had some this morning? Actually. So
there is a graffiti artist who has been just absolutely
tearing up downtown LA and I missed seeing one of
his late his I have an idea of who it is,
but I miss seeing one of his latest works by
(00:56):
probably a few hours. It was being painted over. As
somebody told me about it. I went to the location
and somebody's painting over it. So I missed it.
Speaker 1 (01:04):
I missed it, Natasha, anything from you?
Speaker 6 (01:07):
Well, this week was maybe the first time I ever thought, oh,
what if I got to be a White House reporter
and like be in China with Jensen Huang and and
uh Trump and see how those negotiations went, be the
only woman in the room. So yeah, otherwise it doesn't
really appeal to me.
Speaker 1 (01:27):
Well, you said, that the two magic words for this
week's episode intro, which is Jensen Huang and fomo, because
he almost didn't make it, right.
Speaker 3 (01:35):
Rud, That's right. He wasn't going to go. It was
too politically touchy.
Speaker 1 (01:41):
Now it wasn't He wasn't He initially not invited and
then he literally found a way to get to Alaska
and jump on F Force one mid trip I mean
not one plane to plane, not like top gun.
Speaker 3 (01:49):
But yeah, no, it wasn't a mid air transition, although
I feel like with that leather jacket he probably could
do that. But you know, no, it was like if
the President said, hey, I want you to come to
China with me, hop on Air Force one. We're going
to stop in Alaska, and you're a billionaire, you just
you just take your plan to Alaska.
Speaker 2 (02:08):
I guess hop on No big deal.
Speaker 1 (02:11):
Why wasn't he because so Tim Cook and Elona there.
Why was Jensen not originally included in the in the past, Well, he's.
Speaker 3 (02:18):
Been advocating to you know, to sell basically to have
like basically no restrictions on on chip sales to China,
and he's got some of what he wanted, and there's
and there's a conflict there with the China Hawks who
think that's a bad idea. It's also there's just a
ton of scrutiny around him and how close he is
to the president. Meanwhile, AI, you know, AI is becoming
(02:41):
very unpopular politically, and it's it's sort of becoming a liability.
So it was like, Ah, this is going to cause
some awkward conversations. Let's not bring him. And then I
think that became a bit of a news cycle. Maybe
it was on TV and the President saw it and
was like, actually, no, I want that guy. I want
leather jacket guy. Bring him.
Speaker 1 (03:02):
Let's get into it. Welcome to Tech Stuff. I was
Volosha and this is the Week in Tech where I'm
joined by three of the most plugged in tech reporters
to break down what's really happening right now. Today we're
joined by Reid Albergotti from Semophore pul it, a prize
winning journalist, Dexter Thomas, and Natasha Tiku, tech reporter for
the Washington Post. Read I want to start with you.
You told us on this show back in March about
(03:24):
this brewing battle between the AI companies who has most
compute and then recently wrote about how Elon Musk's space
Xai is licensing its computing power to Anthropic. So I
want to talk about that deal. But before we get there,
you've been you've been token maxing on this show for
(03:44):
a while, and I'm just wondering if you could give us,
give us your explanation of why why we should all
be paying as much attention to the new token economy
as you are.
Speaker 2 (03:54):
Yeah, I mean, yeah, I don't.
Speaker 3 (03:55):
I don't think i'm token maxing compared to you know,
a lot of people out here in the day area.
But I do use these things every day, and I'm
finding I'm getting a lot of use out of them.
And I'm also sort of aware, just based on my
own informal polling that you know, I'm actually a pretty
early on the adoption curve here. Like I don't, I
don't think we've really seen anything yet in terms of
(04:15):
how many tokens people are going to be using when
when this stuff really catches on and gets better and
easier to use. And you know that just to me
says a lot about this token economy, like the demand
is going to be there for a long time.
Speaker 1 (04:29):
But I want to ask you about that specifically because
you've made this argument quite you know, consistently for the
last few weeks. I haven't really seen anywhere else and
we haven't actually paused to unpack about tokens as a
kind of new currency or commodity, Like are they like bitcoin?
Are they like US dollars? They like gold? Are they
like cloth? Are they like you know, oil? Like what
is the what is the most familiar form to understand tokens?
Speaker 3 (04:51):
Yeah, I mean they're kind of like bitcoin, but they
actually have utility, right like people people need them and
want them, and you sort of realize that when you
try to use you know, the endthrop products, because you
run out of tokens pretty fast, and then you find yourself,
you know at qwo am sort of like a heroin addict,
like just wanting more tokens, you know, And I think
I think that's like to be really indicative. And and
(05:13):
this elon the whole XAI or SpaceX, whatever you want
to call it, deal with Anthropic where you know, XAI
found itself actually with some extra tokens, and instead of
that being a problem, they just unloaded those two Anthropic
which was short of tokens, and that to me was like,
oh wow, like these things are if their general purpose
they're transferable, they are. They are really becoming like a
(05:37):
like a commodity that you can just easily crede.
Speaker 2 (05:40):
Just fascinating.
Speaker 1 (05:41):
I want to I want to come to Dexter, who
is chuckling about about you shooting up tokens late at night?
Just just before that, really I had one question for you,
which is which has been burning on my mind actually,
which is how much surplus value is there in token creation?
I mean, in other words, like today, when you assess
the price of building a data center, the extric see
that goes into the data center, can minting tokens be
(06:03):
a profitable endeavor or the tokens cost more to make
than their worth today?
Speaker 3 (06:08):
Well, I think it's hard to I mean, these companies
are losing money, so you might say, oh, well they're
actually you know, they're actually not. It's not a profitable business.
But you know, I don't, I don't exactly. I mean
there's a lot of I don't exactly know where the
all the costs are in that process. But we do
know that with compute, it's just it always the price
(06:28):
always comes down. That's just how it works. And you know,
I don't. I don't think we have we want to
spend the whole show like talking about about going into
like the weeds of that. But I think it will.
It will become a very profitable business. But there's also
like a curve where while the demand is like extremely high,
I think it's going to be very very profitable, and
(06:51):
Xai can say, oh, yeah, like this is this is great,
like we'll just unload our our tokens and it works
for everyone. They can do that now, but eventually this
is going to become a mature business, and just creating
the tokens will be a profitable, but like a much
smaller margin business than if you're offering a service on
(07:11):
top of that. So nobody wants to be just a
dumb token provider. That's not what I'm saying at all.
But but I think for right now, the demand is
so high that it actually is a pretty pretty good business.
Speaker 1 (07:21):
To text to you chuckling about about reads a description.
What's what's been your interaction with token mixing to date?
Speaker 5 (07:27):
Oh my god? Well, I mean have we all heard
about the Caveman token hack or the Chinese token hack? Okay,
all right, I'm very sad to introduce this to the group,
But basically, this is a It's really interesting that most
in general, most people don't necessarily know how the inner
workings of AI works. But we're starting to see just
(07:50):
end users understand this to a degree, right right, And
people have come up with these kind of like chat
gbt slash claude life hacks. People are speaking to chatbots
and having the chatbots speak to them as briefly as
possible and just telling it be as terse as possible.
(08:11):
And the absolute most terse you can be is what
people have called like Caveman mode, which is don't use
any don't use prepositions, just say job done, need fix
like things like that, and you speak to the chatbot
like that it speaks to you. But it also just
so happens that English is kind of an inefficient language
(08:31):
because we have a whole bunch of letters that don't
actually do much. Chinese is much more efficient in that
it's able to pack a lot more meaning in, for example,
just one character, and so Chinese users have realized, wait
a second, I can just why would I talk to
this thing in English? I can just talk to it
in Chinese. The problem is It doesn't necessarily always work
as well as you might hope, but you know, people
(08:53):
are hoping to eke out an extra ten, maybe fifteen
eighteen percent in terms of of token usage by either
speaking like a caveman uh and saying good job, bad job,
please fix, or speaking a literal other language so they
can not be sitting there at two am waiting for
their Claud tokens of reload. Maybe they're sitting there at
(09:15):
three am waiting for the Claud tokens a reload.
Speaker 1 (09:18):
But there's an interesting there's interesting cultural tension here because
on the one hand, if you're paying for your tokens,
you want to be in caveman mode and not burned
through too many. But if you work for a tech
company like I guess that you want to yeah, but
precisely building the most inefficient possible ways of working to
get through as many tokens as possible yea, which is
just wild to think about.
Speaker 2 (09:40):
In my experience.
Speaker 3 (09:41):
That's not actually that doesn't actually end up saving you
that much that many tokens, because it's like you're then
just having to do more, you know, more prompts to
get to the you know, you want to be like
as detailed and as specific as possible when you're starting
out in a project. But it is an interesting one.
And like, you know, I had an experience. I built
a website the other day and I was sending it
(10:02):
around to friends and each time they use the website,
it was costing me money, right, And so I had
it build the dashboard to see how much money this
website was costing me, and it was like about seventeen
cents every time someone was using it.
Speaker 2 (10:16):
And so then I had it.
Speaker 3 (10:17):
I was like, just run an evaluation and you know,
find a better, more efficient way to do this. Here's
a bunch of different APIs, try them all out, and
it brought the seventeen cents down to like a fraction
of a penny and it still works.
Speaker 2 (10:31):
And I was like, oh, that's pretty cool. You know,
there's very.
Speaker 1 (10:33):
Interesting I got. I got LinkedIn shamed. I wish there
was a regular experience by somebody telling me that as
a founder, if I wasn't spending as much a month
on tokens as I am on rent, I'm already doing
myself a huge disservice. So goals.
Speaker 5 (10:48):
This is the really interesting thing, though, is I mean
you brought up bitcoin, which obviously this is not a
one to one comparison. But the really interesting thing is
there are people who were really really into bitcoin and
they don't think or ethereum, but whatever cryptocurrency you want
to talk about. They don't think of their net worth
in terms of dollars. They think of their net worth
(11:09):
in terms of bitcoin. How much bitcoin.
Speaker 1 (11:11):
Doner, the conversion is not they're not. They're not. They're
not even monitoring the exchange rate.
Speaker 5 (11:14):
Or they'll tell you that they're not right. They are.
They know exactly, they know that they know the Bitcoin
to US D ratio, they know what it is. But
they'll tell you. Don't think of your money in terms
of fiat. That money is dirty, it's the government money.
Think of it in terms of bitcoin. And it's very
interesting that you're having people think of tokens as a
(11:35):
unit of value. That that to means fascinating.
Speaker 3 (11:38):
I just think of how much I owe my mortgage
that's being in California. But go ahead, I didn't mean.
Speaker 1 (11:45):
Well, Nontas, I just wanted to say to you. I mean, obviously,
the prompt for this conversation, so to speak, was this
was this space Xai anthropic deal which kind of interesting
right space XAI. Elon was praised for building these massive
data centers early, and everyone was jealous and apparently like
other companies were flying drones over his building sites because
he was like in maniac mode and doing construction much
(12:07):
fast than everybody else. But then, of course his product
isn't like as useful in terms of actual AI than
Anthropic or Open ais Le Google, So it's kind of
interesting irony there. But also I think three months ago
Elon said that Anthropic was misanthropic and it was evil,
and there's a break clause in this contract if he
decides to Anthropic it evil again. So I mean, do
(12:29):
you see another lawsuit brewing here? Like what hell is
this going to go down?
Speaker 6 (12:34):
I I mean it just it just goes to show
you like Compute brings all men together, you know, they
like when it's when it's a deal of convenience. Elon
conveniently changed his mind about who's okay being in charge
of AGI. I just find his like word play to
(12:55):
be so bizarre, like why fixate on the name of
open AI being open or misanthropic, Like it's just.
Speaker 1 (13:07):
It's quite trumpy and right, It's like a very very yeah.
But I mean it's actually been reporting on Silicon Valley
for a while. I remember when like when you know, Google, Facebook,
you know, et cetera, were and Apple were kind of
all becoming behemoths in like twenty ten era. Eric Schmidt
(13:28):
was on Apple's board i think, and there was quite
a lot of like sort of circular. It's a small community,
and it felt like a rising tide was, you know,
raising all boats, and then all hell broke loose here.
It seems like it's not sequential like arising. Everyone feels
like a rising tide is raising all boats and they'll
work together, but all hell but it's also breaking loose simultaneously,
which is a kind of an interesting conundrum.
Speaker 6 (13:49):
Yeah, it's a tangle of ties, but tenuous ties that
could you know. Also, you know, we're all we're all
living in a billionaire's world where they're ego you know,
even sometimes more so than their commercial needs, are dictating
our government policy, our geopolitical rivalries. You know, all of
(14:15):
a sudden, Trump might come back, like depending on what
happens with Jensen Wang and you know in China, might
come back saying like, oh, you know, she's our buddy, Now,
everything's fine. He was already moving away from you know,
talk at least about selling weapons to Taiwan and protecting Taiwan.
(14:35):
So truly like five to ten men control the world.
Speaker 1 (14:41):
Yeah, I mean we're going we're going to talk about
Chavon Jis after the break. But it was interesting. He
brought three hardware people, right, I mean, Tim Cook, crom Apple,
Jenson Huang from a video you know, must from TESTA
and SpaceX and others. But but none of them. But
neither Samilman nor Dariama Day nor you know, the kind
of fundamental AI model we're not included in the in the.
Speaker 6 (15:02):
Picnic, Yeah, he had some you know, like more mature,
grown up, regular business folks like that. I'm sure he
sees more often in New York. But I don't think
you're going to bring Darya Amiday, Considering you know, he
is a huge China hawk and his rhetoric around their
(15:24):
authoritarian government. I think he would definitely be off the list.
Speaker 1 (15:28):
We're going to take a short break now. When we
come back, Atash will give us the latest updates from
the Elon Open Ai trial and read you're off to
shoot some peptides.
Speaker 2 (15:37):
Yes, I am good luck.
Speaker 1 (15:45):
When you're traveling abroad, wi Fi is always an issue.
Maybe you can't find it, or maybe you can, but
the Wi Fi network that's available looks a little bit
sketchy and makes you think it might not be safe
to connect to. I've certainly had that feeling, and it's
why having a local simcard on your phone can make
all the difference. But you don't have to literally buy one.
(16:06):
That's where Sale comes in, an e SIM service that
is as simple as downloading an app, and that gives
you instant Internet access wherever you're traveling. That is safe
and secure Internet with built in cybersecurity features like web
protection and ad blockers, and no need for local Wi Fi.
All you have to do is download the Sale app
on your mobile device and choose the plan that works
(16:26):
best for you. Once you install the e sim, Sale
will be instantly activated, so you can start your trip
off right. Download salely in your app store and use
our code tech Stuff at checkout to get an exclusive
fifteen percent off your first purchase. See details in the
podcast episode description box. Welcome back to tech Stuff, Natasha.
(16:47):
Let's move onto your story. You've been filling us in
on the details of the Elon Musk open Ai trial
remind us what the trial is about and why Shivon
Gillis is such an important character.
Speaker 6 (16:57):
So this trial is about, as Elon must put it,
you know, did open ai steal a charity? Did they
take this nonprofit company which was supposed to be you know,
a counter force against Google or China or you know,
some other large entity having control of ai in the future.
(17:21):
Did the idea of them getting money from Microsoft, you know,
getting billions of dollars from Microsoft? Did that change the dynamics?
Did Sam Altman, you know, enrich himself at the cost
of humanity? Is basically Elon's argument. So what's central to
that are the negotiations that the board had with changing
(17:42):
the structure of the company from a nonprofit to a
for profit. And Chavon Zillis was on the board at
this time. You know, she approved those decisions to get
funding from Microsoft. So she is, in a way like
a key witness for open Ai for the defense.
Speaker 1 (18:03):
Which is interesting. So she's she's the mother of four
of his children. She was on the board of open Ai,
but was also communicating with Elon during that time. But
now she's essentially a witness for open AI. Was she
called by by the defense or how did she end
up on the stand or what was it? Like what
happened there and she.
Speaker 6 (18:22):
Ended up on the stand because so much of her
communication was you know, in the discovery for this trial,
many of her notes, her personal text messages. They're clearly
very close and clearly her loyalties which came out in
her testimony, Clearly her loyalties are towards Elon Musk.
Speaker 1 (18:40):
I want to ask you more about about the trial
and what's come out during the during this her on
the stand, but you know you had this phrase before
the break about what was it tokens bring all men together?
What was it? I mean, this is compute brings all
the men together. I mean this is interesting. She's a
female board member. What are the kind of gender dynamics
of hivon the bros.
Speaker 6 (19:01):
Yeah, I mean you see her, this is like the
one woman right like that is that has become a
central figure in this trial. And what she's doing is
essentially kind of you know, in some cases like massaging egos,
smoothing tensions, you know, trying to translate one person's needs
(19:23):
for another. Like I said, like emotional labor for these men,
and you know, it's just it's kind of it's frustrating
to watch. There's one exchange where sam Altman is asking her,
is she going to attend, you know, one of the
rocket events, and she said, oh, no, no, no, no, I
wouldn't even dare ask. And it just, you know, this
(19:47):
deference that she has towards Musk, you know, we just it.
I guess it makes me wonder, you know, how she
describes some of these things to her friends or to
her family, if she does it all. We also learned
on the stand that her until until Business Insider reported
(20:08):
that Musk was the father of her children, even her
own father did not know that.
Speaker 1 (20:12):
In fact, I've read that her first call when they
called her to fact check the story was to her
father and her second was dissembleman, Yeah, what a life.
Speaker 6 (20:20):
I know. I mean, you know, this is a woman
who has tied her future, her personal life, her fortunes
to a very mercurial man who describes her as as
his chief of staff. And sorry, just to clarify too, Chavon.
At the same time that she was on the Open
AI board, she was advising Musque on neuralink Tesla, you know,
(20:43):
his kind of his portfolio across the board.
Speaker 1 (20:46):
Texter, what's your take on with this?
Speaker 5 (20:48):
I keep thinking of, you know, how so many people
think that their phone is listening to them, and yet
they use their phone anyway. And I think a lot
of people are watching, you know, if they're paying attention
to this at all, the you know, this sort of
courtroom drama thing happening and feeling like, well, they go
at it again. It's as bad as I thought it was,
(21:10):
but here's a few extra details. It's it's sort of strange.
It's strange to watch.
Speaker 6 (21:16):
Yeah, I think if you if you pull back a
little bit too. Part of a large part of what
this trial is is a number of people defining whether
Open AI is adhering to its mission, which was building
AI to benefit all of humanity. And if nothing else,
(21:36):
you get absolute confirmation that it is a handful of
people that are, you know, having their personal perspective on
whether this is good for humanity or not.
Speaker 1 (21:47):
This is the.
Speaker 6 (21:48):
Least democratic process, you know, you have ever seen. There
is one exchange where Shavn is trying to think about
how what are different solutions for getting the money that
need you know, that's not working as a nonprofit. And
she suggests talking to demissabas, the Zee of Google Deep Mind,
(22:08):
and she describes him, you know, like whether or not
he is moral enough, and she says, oh, maybe if
he spends more time around e he'll start thinking about
humanity more. And yet all of the interactions we're seeing
from Elon are concerned with whether or not he's in
control of the company. It's just it, just like brings
(22:30):
the ridiculousness of open AI's stated mission into clear focus.
Whether or not your entire court case revolves around protecting humanity.
These are not people who are talking about humanity. These
are people who are talking about whether or not they
exert control over a product and a technology deck.
Speaker 1 (22:51):
So you put that parenthetical if they're following it at
all in your coined about the time, what do you think?
Speaker 2 (22:56):
Is how nice?
Speaker 1 (22:57):
Hard to know? When you I mean, you know, we
and most of my time reading tech news, right, so
I have been following it, But I don't know. And
this is obviously not the oj trial. But is it
is it?
Speaker 2 (23:07):
Is it?
Speaker 1 (23:07):
Is it something which has permeated the popular consciousness? Do
you think?
Speaker 5 (23:10):
I don't think so, I don't think so, I think,
I think there's a section of us who are definitely
paying attention. But I think, you know, I have conversations
regularly where I'm I'm talking about some of my work
and I say Sam Altman, and I get a who
Who's that? Who's that? And so obviously not everybody's tapped
(23:30):
in with this. I think they. I would hope that
more people would. And I think, Natasha, precisely, what you're
talking about and what's being discussed in this court case
is something that you see and un frankly, a lot
of Sam Altman's projects, I mean, look at one of them,
world Coin, right, which is now Tools for Humanity, which
is you know these orbs right. One of these stated
(23:51):
purposes originally of one of one of his projects was
to provide ubi universal basic income. So AI is going
to take everybody's jobs, and so in order to prevent
the damage or you know, mitigate the damage it will
come from that, let's make sure that everybody in the
world is getting, you know, a basic income. And of
(24:15):
course they started by going to places like Kenya and
scanning everybody's eyeballs. But now if you look at the
language on the website and when they talk about it,
they're not really talking about universal basic income anymore. They're
talking about verifying that you're a human. And if you
keep pressing them about why is that important, eventually you
start hearing about, well, advertisers want to make sure that
(24:38):
the people who are seeing their ads aren't bots, they're
actual humans, even if you know. So, there's a lot
of this shifting, saying one thing and then making sure
that the public thinks, oh, yes, this is what they do,
and then what's actually happening starts to change. And I
would love for more people to you know, be tapped
(24:59):
in aware of this. That is one of the purposes
of the show, of course, but it's tough. It's tough
because it moves so fast. It moves so fast.
Speaker 6 (25:08):
Well, I think I have started to see, at least
with Sam Altman. I think partly because you know, the
nature of how we consume tech information is we treat,
you know, the CEO as like a proxy almost for
the company, and we have this impression that they are
controlling things, often because they do have you know, super shares,
(25:30):
they tip the board in their favor, et cetera. So
I do you know, yes, nobody I think no regular
person is paying attention to the trial, But I think
people do want to get a read on Sam Altman.
I think they do, you know, when the Pentagon stuff started.
They do want to get a read on Dario Amiday.
And with good reason, right like, as you so rightly
(25:52):
pointed out, Dexter, Like what they say is is is
never ends up being like a good indicator of how
their technology is going to impact the public. So you
do need to figure out, you know, how they operate
as a person, you know, what they really care about,
and the language from open AI since the beginning, you know,
they're almost flirting with talking about like wealth redistribution. You know,
(26:18):
they'll like the way that they talk about human flourishing,
or you know they kind of wave at what will
happen when all the jobs are taken away. I think
I think the reason that people are grasping for psychological
reads sometimes on these CEOs is because they know it
will you know, have economic consequences for them in the.
Speaker 5 (26:41):
Future, definitely. And I think the the other thing that
I've been trying to puzzle out, like why is it
that people don't know the names like Amada, Why don't
people know Sam Altman. Some of it is because this
has happened so fast. Like open AI fundamentally for most
people did not exist three years ago. It wasn't a
function anthropic even more so. Right, But I don't think
(27:06):
it's just that. I think, you know, people can name
Tim Cook, people can name Steve Jobs. Of course, people
can you know, go back a little bit further, you know,
Bill Gates, all these sorts of things.
Speaker 1 (27:16):
The thing that Simulman has higher name recognition than Tim
Cook today in America. No, I don't think interesting. I
don't think Natasha.
Speaker 6 (27:24):
I think so really, I all the time, you know,
we have to gauge this because we like fight over
whether or not we can put his name in a
headline for the Post all the time. And I think
he crossed over partly because he does massive media tours,
you know, like well.
Speaker 1 (27:45):
I think made the Sore app which he was the
main star of as well. Right, I mean, you can't
avoid him in April last year.
Speaker 6 (27:51):
And if you you know, like because there are a
number of influencers now who are just sharing clips of
what these guys say. It's just like, you know, nine
hundred days since you know, one of these AI CEOs
doesn't say something insane, so the clips are shared around routinely.
I think it is just a matter of time before
(28:13):
he becomes a Mark Zuckerberg like figure. I mean, maybe
we need a social network movie.
Speaker 1 (28:19):
But I think I think he became co identified with
AI and like his thing is called open AI, not
anthropic or deep mind, right, and the Chatchipet was the
first product. So I would say people don't probably know
dari Amidau's name or Demos Hassebus his name. But I
my uninformed take is the people that Sam Wlman has
become a kind of public figure, and I think, you know,
(28:40):
the fire bombing of his house as well kind of
goes to show the extent to which he's become like
a sort of figurehead or you know, representative of the
whole industry as well as.
Speaker 2 (28:51):
Just a person.
Speaker 1 (28:54):
We're going to get a break now. When we come back,
we're going to hear about a hacker group who stole
data from over nine thousand colleges and universities, affecting two
hundred and seventy five million users. These hackers wanted money
in return for not releasing the data. Stay with us,
(29:18):
Welcome back, Dexster. You brought a story this week about
the recent hacking of an education platform called Canvas. Yeah,
what happened to you break it down from the beginning.
Speaker 5 (29:27):
So Canvas, with those who don't know, is what's called
an LMS, so I think a learning management system. And
so basically it's the way that a student interfaces with
their classroom digitally. So a teacher might you or a
professor might use it to assign homework, to host group discussions,
to give tests, and maybe most importantly, or to post
(29:50):
readings things like that. Maybe most importantly we're going to
talk about this. You can send messages to your professor
via essentially this portal, and so.
Speaker 1 (30:00):
Slack at tech slack kind of kind.
Speaker 5 (30:02):
Of Yeah, it's a bit like that for those of
you Slack. And so this is if you know somebody
who is going to university or college in North America anywhere,
or you know somebody who has been to college in
the last decade, has some change, there's a very high
chance that they've used this. And I say that because
(30:23):
Canvas holds about forty percent of the market in North America.
Speaker 1 (30:29):
So this is a that's a huge two hundred and
seventy five million number.
Speaker 5 (30:32):
Comes from precisely. So let's talk about the hack. On
April twenty ninth, Canvas was first hacked, and so the
basically the company that runs Canvas is called Instructure. So
April twenty ninth, this group called Shiny Hunters, which is
a Pokemon reference, but they're also very well known, pretty
(30:52):
notorious for hitting a whole bunch of other companies very
successfully and very lucratively. I might add to the tunes
of millions of dollars they hit group. They're all over
the place. They're all over the place. So, and this
is the thing about hacker groups. I mean, if we
knew where they were, it'd be much easier. They probably
wouldn't be operating anymore. But so April twenty ninth they
(31:17):
get hacked. May seventh they get hacked again with the
same method. So, without getting too much into the weeds here,
this is a little bit like breaking into the front
door and then a little over a week later breaking
into the side gate. Like they basically did the same thing,
(31:37):
but the difference is on May seventh, basically they posted
a ransom note on the Canvas portal for all of
these students, So students who were trying to log in
to turn in assignments, and mind you, this is final
season at Harvard, at Penn At a lot of really
(31:59):
prominent elite universities get on and they see a ransom
note saying that hey, we have taken down canvas, We've
gotten all the messages, we've gotten all this other information,
and Instructure won't even talk to us.
Speaker 1 (32:15):
So the first breach, the front door, as it were,
was to basically extract all of the data.
Speaker 5 (32:20):
It looks like they got some, and they may have
gotten more.
Speaker 1 (32:24):
It's a little hard hard to tell, but enough to
be able to do the second breach, to say, if
you don't pay us, we're going to release all this data.
Speaker 5 (32:31):
Is interesting, precisely, precisely. Yeah, And so that's when it's
public in the sense of instead of just talking to
Instructure and saying, hey, give us money, now, they're going
directly to the students, and that guarantees that it gets
media attention.
Speaker 1 (32:50):
And they paid up right, I mean, the ransom note worked.
Speaker 5 (32:52):
Yeah. So okay, the answer to the short answer to
your question is yes, Instructure has posted a message on
the website saying that they have reached an agreement. They
didn't explicitly say that they paid the ransom, but we
know what that means. And they also have said that
the stolen data was given back and that they have
(33:15):
received digital confirmation of data destruction, and they say in
parentheses shred logs. I see your face. I'd like you
to say what you're thinking.
Speaker 1 (33:25):
I mean that sounds like BS to me, right. I
mean the statement they gave, by the way, was, while
there is never complete certainty when dealing with cyber criminals,
we believe it was important to take every step within
our control to give a customers additional peace of mind
to the extent possible. I mean, that is one of
the most overqualified, absurd corporate statements I've had in my life.
Speaker 5 (33:48):
There's a lot of caveats in there. Yeah.
Speaker 1 (33:50):
So, but maybe I'm wrong, and maybe I'm being too cynical.
You think maybe they did credibly destroy the data.
Speaker 5 (33:56):
Data is not like your mother's fine china. It's not
something that you pick up and then you take out
of the house and then you give back. It's Anybody
who's ever used a computer before knows that you can
just make a copy of things. That's what computers are for.
So for the tree, yeah, shred logs. A shred log
is just it's basically a text file saying we deleted this,
(34:18):
but if you made copies of it before. There's no
third party audit, there's no forensic trail. There is a
trust me bro from a group of criminals. This is
how they operate. What we have received is not evidence
that all data has been quote unquote given back or deleted.
(34:40):
What we've received is a pinky promise.
Speaker 1 (34:43):
Natasha, I want to ask you about this because I mean,
this is a story which is affected. You know, a
pinky promise is now guaranteeing the security of the data
of two hundred and seventy five million people, basically everyone
in college and everyone who's been at college in the
last ten years, at all of their professors. I mean,
that is functionally like a huge group. And yet I'm
glad you brought this story. Dexter It got I think
(35:04):
four paragraphs in the New York Times and their longer
piece in the Higher Education Journal. But I would say,
broadly speaking, like no pickup. Whereas mythos, the anthropic model
that has the potential to render all cybersecurity mood is
still dominating the headlines. A month on why Natasha, do
you think that the theoretical power of mythos is so
much more interesting to people than the literal extraction and
(35:27):
theft of this huge amount of data.
Speaker 5 (35:30):
I mean, that's.
Speaker 6 (35:32):
That's such a good contrast, right, I think about this
all the time with with like what counts as AI.
You know, there is obviously the media has a bias
towards what is new, and I think there's just a
sense of powerlessness about how our data is used by corporations.
(35:54):
You know, it's it's not sexy, it's not interesting, it's
just you know, everywhere. Then you have to start thinking
about like ad networks and like Dexter is talking about,
you know, how data gets shared everywhere. You know, you
could also be looking if you cared about AI at
like algorithmic bias in banking, you know, in housing, in education,
(36:19):
et cetera. And yet like that's the focus is on
usually a set of five men. Sorry, keep coming back
to the same theme, but that's the way regulators think,
that's the way the media thinks, and so that is
where our attention goes.
Speaker 1 (36:36):
I mean, Dexter, I think part of the big part
of your your body of work is surfacing stories exactly
like this one that are not the kind of a
one New York Times story about who's up who's down
in the power stakes. I mean, what what did this
story say to you?
Speaker 5 (36:50):
Well, I mean, I'll be real with you, this is
actually pretty personal for me because I teach at universities,
and so, just to be clear, the information that they
got was usernames, email addresses, course names, enrollment information, and messages.
(37:11):
That's the most important part right there, the messages. Right So,
Instructure has come out and said, oh, well, you know,
they didn't. None of the course content was compromised. Nobody
cares if homework number seven from Engineering one oh one
gets leaked on the dark web. Nobody cares anyway, Yeah,
nobody cares about that. The crucial thing here is the messages.
(37:35):
Students have extremely vulnerable conversations with their professors, and it
could be anything. It could be my grandmother passed away
and that's why I can't make it into class. It
can be I'm really struggling mentally and I don't know
what to do, and that's why I haven't been in
(37:55):
class three weeks. You know, can we talk? It can
be this really terrible thing happened to me? What do
I do? I've received all those messages, these are numerous times.
These are at every institution that I've taught at, including UCLA,
which is one of the schools that was hit is.
So that's me thinking as a professor here. Now let
(38:19):
me think as a hacker. The hackers have said that
they are going they are not going to resell the data.
There's no agreement to not use the data. I can
tell you exactly what I'm going to do, what I
would do right now. So one of the schools hit
was Harvard. So I'm a member of Shiny Hunters. Let's
say let's say that I'm not going to resell this data.
(38:39):
Shiny Hunters has resold data before. This is something they do.
It's part of their business model. By the way, the
FBI tells ransomware victims to not pay the ransom for
specifically uses like this because A identifies your company as
a mark and B lets people know that this is
a payday. And this is how shiny Hunters operates.
Speaker 2 (39:01):
Right.
Speaker 5 (39:01):
So okay, let me stop thinking as a professor. Let
me start thinking as a hacker. What I would be
doing right now. Let's say that I'm going to abide
by my pinky promise to not resell this stuff to anybody. Okay,
I'll just use it for my personal use. Here's what
I do. Harvard. I know there's a bunch of rich
kids go to Harvard. I have the contents of the messages,
and I have the enrollment information. So let's say I
(39:23):
just grab a list of a thousand students. I know
what classes they're in, and I know which student is
in which class. So I just send off an email
to a thousand Harvard kids and I say, hey, it's
your bio professor. It's Wednesday. You forgot to sign the
student contract at the beginning of the course. I don't
know how I didn't notice this, but if you don't
(39:44):
sign this by seven pm. By the way, it's five pm,
then your grade is going to be registered as an
F and I won't be able to fix it until
the start of summer semester. So you're gonna have F.
You're gonna have an F on your transcript. I know
that you were going for that internship. I know this
because of the messages. So here's here's an email. Just
click this link, log in, fix it, and boom, phishing attempt.
(40:04):
I got your email address, now I log in. Now
I'm in your iCloud. Now I got all your messages,
Now I got all your photos. Imagine what kind of
photos on there. No, this is and I'm not giving
anybody instructions because this is precisely what I would be
doing right now.
Speaker 2 (40:19):
Yeah.
Speaker 5 (40:20):
No, I know. It is one of those things like
I can definitely imagine somebody saying, Dexter, don't don't give
people ideas. No, no, I'm not giving anybody ideas. This
this is what's happening actually right now. Yeah, fire off
a quick message. Out of a thousand kids, maybe three
get caught, pretty good ratio. Fire off a message to
the kid and to the parents, say hey, I'm in
your kid's iicloud two and fifty K. If I don't
(40:41):
have it by Thursday, it goes up to half a mill.
And you could say, oh, well, I don't care. Harvard.
It's a bunch of rich kids, a bunch of you know,
legacy kids. I do know a bunch of poor kids
who go to Harvard. People get scholarships. So somebody is
going to have their life ruined. Multiple somebody are going
to have their life ruined. Remember, they got billions of messages.
They have a lot of information which is going to
(41:03):
be very useful for social engineering, the best kind of hacking.
And and we will never know because unlike Construction, which
is a public and traded company who has to at
least say, hey, something happened to us. We know that
the psychology of this is people are ashamed when they
(41:25):
get scammed, and so they keep it quiet. And so
we will be seeing We will be feeling the ripple
effects of this for years because if I don't, let's say,
I don't get any kids on that first sweep of
a thousand people, try again in three months. The ROI
on this is astronomically great. We will be feeling it.
(41:47):
We'll never see it because it will never get reported.
I've heard I've heard this said. I'm still reporting on this.
But I've heard this called the worst privacy disaster in
the history of educational tech, and I fully agree with that.
There's been I can't think of anything worse. This is
catastrophically bad. If you know anybody who is a student,
(42:10):
or you know anybody who's a professor, at the very
very least, I know you didn't ask this, But I
feel like I got to say this because I just
said a bunch of bad things. I feel like I
got to give you something. If you know anybody who's
a student, tell them change if they happen to use
the same password in their Canvas account as they use
anywhere else, for example, their email, change that right now.
(42:32):
The next thing would be talk to your university and say, hey,
what are y'all doing? Because there's some things universities can do,
and every university is different, every college is different. But
at the very very very least, it's easy to feel
powerless now, and I f really understand that. But at
the very least, if you're sharing passwords across multiple accounts,
definitely go and change that right now.
Speaker 1 (42:54):
Like now, well, that's the way I have time for
this week. We are going to go out on our
final cake a dexter I think you'll us week in
Tech goes to the studenty student and professor in America. Yes,
my worst week in Tech is also higher ed coded,
but quite a lot less serious and somber than yours.
Gloria Caulfield, who was giving a commencement address at the
(43:18):
University of Central Florida, got booed almost off the stage.
Speaker 5 (43:22):
Oh my gosh, yeah reading the room man not a
strong suit apparently of this individual. That was a rough one.
Speaker 2 (43:28):
Let's say we can play the lip.
Speaker 5 (43:30):
The rise of artificial intelligence is the next industrial revolution?
Speaker 2 (43:41):
WHOA?
Speaker 3 (43:45):
What happened?
Speaker 1 (43:48):
Okay, I struck a chord? May I finish? Thanks for
sound that you'd heard this already. But but Natasha, did
you see this one?
Speaker 3 (43:59):
Oh?
Speaker 4 (43:59):
Yes, yeah, I loved how she like when they started booing,
she looked back at the days like, it's just I
don't know. It was just a very like physical encapsulation
of how Voss layer CEO layer management layer works.
Speaker 2 (44:16):
Like the turning around me, like, what's going on here?
Speaker 1 (44:18):
So I know they're in front of you, that's doing
you because of what you said.
Speaker 5 (44:21):
It's it's so yeah, it's so unfortunate. I mean, I
can't confirm. I have no evidence here, but I wouldn't
be surprised if a few tokens were expended writing that speech.
It was a unfortunately sadly out of touch speech and
it went downhill from there.
Speaker 1 (44:39):
Natsher's is chavon you your West Weekend Tech? What do
you say?
Speaker 6 (44:43):
I'm maybe gonna put Musk in my worst week in tech. Actually,
you know, first of all, I think that there, you
know it wasn't covered this way, but the anthropics SpaceX deal.
Giving up this compute that you knew you would need
or you believed you would need because you thought you
are going to be leading the technology, giving that to
(45:05):
a guy you called misanthropic. That's that's definitely down in
the rankings.
Speaker 1 (45:10):
Okay, let's do our best. One sentence only best for
Jensen Hwang for me because he managed to he managed
to make his dreams come true and get on the
trip after all.
Speaker 6 (45:19):
Jensen for sure. And whoever makes Jensen's leather.
Speaker 5 (45:22):
Jackets everybody who is in the audience at that unfortunate
commencement speech because they're they're famous and they can do
something with it. I don't know what do somebody?
Speaker 1 (45:34):
That's it for the we can take. Thank you all
so much for joining us today the Tech Stuff Finals
Veloci and This episode was produced by Eliza Dennis and
Melissa Slaughter. Executive produced by me Julian Nutta and Kate
Osborne for Geidoscope and Katrian Novel for iHeart Podcasts. The
engineer is Kathleen Conti from CDM Studios. Jack Intolely mixed
this episode and Kyle Murdoch wrote our theme song special
(45:54):
thank you to Read Albergotti, Dexa Thomas, and Natasha Tiku
with Practical and Friends of the Paul. And we'll go
for to see you next week
TechStuff News
Hosts And Creators
Oz Woloshyn
Karah Preiss
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Hey Jonas!
Hey Jonas! The official Jonas Brothers podcast. Hosted by Kevin, Joe, and Nick Jonas. It’s the Jonas Brothers you know... musicians, actors, and well, yes, brothers. Now, they’re sharing another side of themselves in the playful, intimate, and irreverent way only they can. Spend time with the Jonas Brothers here and stay a little bit longer for deep conversations like never before.
Crime Junkie
Does hearing about a true crime case always leave you scouring the internet for the truth behind the story? Dive into your next mystery with Crime Junkie. Every Monday, join your host Ashley Flowers as she unravels all the details of infamous and underreported true crime cases with her best friend Brit Prawat. From cold cases to missing persons and heroes in our community who seek justice, Crime Junkie is your destination for theories and stories you won’t hear anywhere else. Whether you're a seasoned true crime enthusiast or new to the genre, you'll find yourself on the edge of your seat awaiting a new episode every Monday. If you can never get enough true crime... Congratulations, you’ve found your people. Follow to join a community of Crime Junkies! Crime Junkie is presented by Audiochuck Media Company.