233 | ChatGPT porn chat is coming, Job quake: 100 million roles at risk this decade + Mechanize’s “total workforce automation”, the biggest lessons from AgentForce, AI cancer-fighting novel breakthrough, and more AI news for Oct 17, 2025 - Leveraging AI

Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker (00:00):
Hello and welcome to a Weekend News episode of the

(00:02):
Leveraging AI Podcast, a podcastthat shares practical, ethical
ways to leverage AI to improveefficiency, grow your business,
and advance your career.
This is Isar Meitis, your host,and like every week we have a
lot to talk about.
We are going to cover three deepdive topics today.
One is going to be the impact ofAI on jobs and the political

(00:24):
aspect of it as it is evolvingahead of the elections that are
just around the corner.
We're going to talk about thenew release by open ai, which is
very interesting and how they'regetting a little complicated
with the direction that they'retaking right now compared to
what they promised the world.
And we are going to finish thedeep dive with sharing what
happened at Agents for at AgentForce this week, which is

(00:47):
Salesforce big event thathappened this week.
And then we have a long list ofreally interesting rapid fire
items, including new releasesfrom Anthropic, from OpenAI,
from Google.
And we're gonna end with somereally good news.
I always like to end on apositive note, so if you don't
have time to listen toeverything, jump to the end to

(01:08):
listen to the final segmentbecause you will find it very
interesting, some hope forhumanity of potentially curing
cancer with ai.
But that's all the way in theend, and we're in the beginning.
So let's get started.
A Senate Democrat report thatwas just released paints a very

(01:31):
dark picture of the AI'spotential to disrupt the US job
market, potentially taking awayalmost a hundred million jobs in
the coming decade.
This is a significantly highernumber than any other projection
that we've seen before.
This report was spearheaded byBernie Sanders and he is

(01:54):
highlighting an urgent need tocreate policies to protect
workers against AI, automationand robotics that he is going to
take their jobs.
Now, the report itself wasactually based on a ChatGPT
driven analysis of various jobsthat exist today and AI's
opportunity to actually automatethem across different types of

(02:15):
ai, including humanoid robotsand other robotics.
And now I'm quoting Sandershimself.
He's saying Artificialintelligence and robotics will
allow corporate America to wipeout tens of millions of decent
paying jobs, cut labor costs,and boost profits.
Now the report argues that AIconcentrates wealth and power

(02:36):
with obviously tech executivesand tech companies, are leading
this charge as they're investingbillions of dollars to reduce
labor costs and boostproductivity.
Now, on the other side of theaisle, republicans are
advocating that US leadership inAI is a necessity for the future
of the dominance of the us andthey're warning that

(02:59):
overregulating, this field willcut our ability to be
competitive and we may lose therace to other companies, mainly
China.
Now the report by itself isobvious.
The direction is very clear.
The numbers are questionablebecause of the way the research
was done.
It's just relying on ChatGPT,analyzing the current thing, but

(03:19):
the direction's very clear.
AI can take jobs.
Yes.
Is it a political aspect Rightnow?
A hundred percent.
There are elections just aroundthe corner.
We are just a few weeks away.
And job loss is a big thing asfar as driving people to change
their direction and how the waythey vote and job loss is always

(03:40):
a very good threat to dangle infront of people, to scare them
into making one vote versus theother.
Now, the interesting thing here,again, to make it as broad as
possible, because this is apolitical play, they have
mentioned several differentindustries that make it very
hard, and many of them are nottech jobs or leadership jobs,

(04:01):
but they mentioned fast food asan industry that's gonna hit
very hard accounting, trucking.
So you can see that many ofthese are blue collar jobs and
not white collar jobs.
And the direction is very clear.
Again, this is showing you thatthe report is trying to be as
broad as possible in its impacton society.
Despite the fact that I think weare a while out from humanoid

(04:25):
robots serving drinks atStarbucks, the fact that they
mention it as a very highlikelihood that this is going to
be dramatically disrupted, Ithink shows that there's a very
clear agenda behind this report.
I'm not saying it's wrong, bythe way.
I think the timeline might be alittle compressed at the end of
the day.
They're talking about a decade.

(04:45):
There's basically gonna be manyother elections and midterm
elections between now and thetime it's starting to be
critical but I'm not saying thatit's totally wrong, meaning I do
believe that AI and roboticswill take a lot of jobs away.
I think the system will have tochange to deal with that.
And we're gonna talk about otherpeople's opinion, but before we
get to other people's opinion, Iwanna share my own opinion on

(05:09):
the topic.
In my opinion is that the stateslash country level of AI
adoption is very similar to thecompany level decision making
that has to happen.
And I will explain.
A lot of people come to me and Ijust came back from MAICON,
which is an AI conference, andit was fantastic, and I got to
have a lot of AI relatedconversations with people.

(05:29):
And because of this podcast, alot of people know me and came
to me to ask me big questions.
Many of them are around job lossand what decisions can company
makes to stop that.
And I don't think we can, andI'll explain what I mean.
And I said that multiple timeson this podcast.
If you are running a company,you cannot decide not to run

(05:50):
fast with ai.
I mean, you can, but what isgoing to happen?
And my answer to these peoplewhen it comes to the impact on
AI on companies and how can theyapproach AI to not cut jobs.
So there are many businessleaders out there who truly,
deeply care about the human.
Aspect of their company whotruly, deeply care about their

(06:12):
employees and want the best foreach and every one of them.
And they don't want to let anyof them go.
And I think this can last arelatively short amount of time.
I mean, caring about them isobviously something that will
last forever, but being able tokeep all of them is not
something that will lastforever.
Because there are two optionsand two options only if you can

(06:34):
grow faster than theefficiencies that you're
getting.
Meaning if you get a 30%efficiency and you can grow your
company by 50% because you freedtime in your employees and
they're doing higher level tasksand you've learned how to really
leverage AI to grow faster thanyour competitors, then go right
ahead, hire more people to dothat.
However, the pie, the total sizeof the pie, the all the goods

(06:57):
and services that we can sell ina specific industry is finite.
Will it grow with ai?
Maybe in some industries?
Definitely not in all of them.
Let's say you're doing plumbing.
The amount of plumbing needed inthe US or in the world is the
same.
I mean, it's growing a littlebit because we're building more
homes, but nothing dramatic.
Nothing changes from whathappened in the past few years.

(07:18):
And so once everybody finishesgoing through this transition,
it will be very hard to takemore market share, take a bigger
chunk of the pie, and thenyou're competing against other
companies who are all using AImaybe more effectively, less
effectively, but you won't beable to continue growing at the
pace that AI will give youbenefits.
And then you have two options.

(07:39):
Your competitors will be able tosell the same goods and services
you are selling at 10 or 20%less while still making more
money because their coststructure, because of AI, is
gonna be significantly betterthan yours.
And in that case, you have twooptions.
Option one is to lose yourentire company because you will

(08:03):
run out of business because yourcompetitors are gonna be
significantly more efficient andprovide better and faster goods
and services than you can.
That means your company is done,and that means you're not taking
care of any of your employees.
The other option is to startcutting people, to stay
competitive and then you cansave some of the employees.

(08:23):
I think the same exact logicapplies at the national level.
If we try to keep all theemployees and stop ai, then
other countries will becomebetter and more efficient than
us in literally anything.
And then we are going to loseeverything the US has built in
the last 300 years, anddefinitely the domination it

(08:45):
develops since the end of WorldWar ii.
And so if you are excited bythis move by Bernie Sanders, you
gotta keep this in mind.
I'm not taking sides here.
I'm not saying these guys areright and these guys are wrong.
I think the right way isprobably somewhere in the
middle, as I believe in many,many different cases when it
comes to politics, becausepoliticians are taking more and

(09:06):
more extreme positions on moreor less every topic.
the reality is very clear, andas I mentioned, the same concept
that applies at the companylevel, applies at the national
level.
And we will get to the pointthat we just don't have a
choice.
Now, what does that mean?
It means, as I mentioned, thatwe'll get to the point that we
don't have a choice, and doesn'tmatter whether you're a

(09:29):
Republican or a Democrat, wewill face that fact.
And yes, it may take threeyears, five years, 10 years, but
we will get to that position andthat means the entire system of
how we live we'll need tochange.
Now, how exactly will it change?
Well, it will change too.
I don't think anybody knows atthis point, but we will have to
figure it out more about thatfrom a slightly different angle

(09:52):
in the next few segments.
Now staying on the impact of AIon the country, there is a
nation backlash growing againstAI data centers in multiple
different locations,highlighting the cost and impact
of these data centers on localcommunities.
So to put things in perspective,a typical data center consumes

(10:14):
electricity that is equivalentto a hundred thousand households
now, in most cases, it doesn'tcome with building new
electrical power plants todeliver the power to an
additional a hundred thousandhouseholds, which means it's
using the existing grid in theexisting power generation.
In some cases it is usinggenerators which have their own

(10:34):
issues because they're doingsignificantly higher pollution
than a large power plant.
Definitely a nuclear powerplant.
Now, the largest data centersare 20 x, that means 2 million
households as far as the totalelectrical power that they're
consuming.
Now.
In addition, they're consuming alot of water.

(10:56):
So in Mansfield, Georgia made usdata centers have left many
residents having to buy bottledwater for drinking because they
are claiming, and again, Ididn't verify that, that the
local whales became murky andthe local water became
undrinkable because of theamount of water that gets

(11:18):
consumed by the data center.
Now, is that, again, a politicalstatement?
How much of the wells becomingmurky is really because of the
data center versus just theoverall population and the
amount of rain and many otheraspects?
I don't know, but it doesn'tmatter because this is the
mindset that people have andthis is what they feel that data
centers are doing.

(11:39):
Looking back at the electricalaspect of it, Mike Jacobs, who
is the senior energy manager atthe Union of Concerned
Scientists.
Again, you can get by the name.
What are they concerned about?
He found that data centers added$1.9 billion to Virginia
households electric bills in2024 alone.

(12:00):
Now, again, I wasn't able tofind exactly how they did the
math, but it doesn't matter.
And again, the reason it doesn'tmatter is it is obvious on one
hand that AI data centers areconsuming a lot of electricity
from the existing grid.
It that means that there's moredemand.
That means that the price aregoing to go up and whether the
number is 1.5 billion or 2million, it doesn't matter

(12:22):
because the direction is clearand that's what people care
about.
And to show you how much thissituation is extreme right now
in Virginia's 30th HouseDistrict, both Republicans and
Democrats campaign against localbuilding of data centers.
And the quote from JohnMcAuliffe, who is the Democrat

(12:42):
that is leading it togetheragain with a partner from the
Republican Party.
He said, and I'm quoting, weneed to make sure Virginians are
benefiting, not just paying forit.
Now, when was the last time yousaw a unified bipartisan support
of anything in the past fewyears?
This just shows you how criticalthis is becoming, again,

(13:05):
especially going into electionsand how much politicians care
about job creation and impact onlives when it comes to basically
everything just before theelections.
And so we're going to hear moreand more about this in the next
few weeks as the election comesup.
Will the volume of this kind ofprotest and so on continue after

(13:27):
the elections?
I'm not a hundred percent sure,but right now there are many,
many fronts in which localorganizations, communities, and
so on, are fighting againstbuilding data centers in their
area now do I think this is ajustified battle by the local
people of these differentcommunities?
I a hundred percent think thebattle is justified.

(13:49):
I do truly believe that datacenters consume a lot of energy
and a lot of water, and thatwill take its toll on local
population.
Do I think the local people havea chance to win these battles?
I think they can win a few.
I am a hundred percent certainthey're going to lose the war.
I'm not saying that being happyabout it.

(14:11):
I'm just saying there's way toomuch money in huge interests
that are involved, and there isa government right now that sees
the race with China as critical.
And so I believe that most ofthese data centers will get
built in the locations that thecompanies want to build them.
Again, good or bad, helping thelocal population by delivering,

(14:34):
uh, jobs, et cetera or bad forthe local population.
I think time will tell, we'll beable to analyze this a few years
into the future, but right now,again, I think other than maybe
a few local battles, I think thebigger companies are going to
win.
Now, continuing on the topic onthe impact of jobs in the US
society and population.
A 16 Z, which is one of the mostknown and successful vCs in

(14:57):
California, and they'reobviously deeply involved in the
AI universe, has released apodcast, their own podcast, a
very interesting episode.
It is a presentation by AlexPel, who is a general partner at
A 16 Z, and he delivered thatat.
On a presentation he did at theLP Summit, and it's titled
Software Is Eating Labor.

(15:19):
Now, first of all, a little biton the title, Mike Anderson, who
is the founder of this vc, hadan essay in 2011 that was called
Why Software Is Eating theWorld.
So this is a homage if you want,on that particular piece.
So this piece is called AI'sEating Labor, and he makes some
very interesting and troublingclaims, but they all make

(15:41):
perfect sense.
So what he mentioned is that thecurrent worldwide SaaS market is
about$300 billion in revenuethat it's generating every
single year.
This is definitely a really bigpile of money for AI to go
after, but what he's also sayingis that the US labor market
alone spends$13 trillion peryear on salaries.

(16:05):
That's a much bigger pile ofmoney.
That's two orders of magnitudebigger.
Now Rampal does an interestinghistory review where he's
basically showing that thegrowth of SaaS company or
software companies in generalhas literally just taking files
and digitizing them.
He gave multiple examples, butjust to name a few.

(16:26):
Sabre, which is still thebackbone to many of the way we
order airlines, which wasdeveloped by IBM.
Was literally digitizing paperslips that previously held all
the information on whichairplane takes off at what time
and who sits in what seat and soon, and made it into a digital
environment, Salesforce, andbefore that on premise, CRMs

(16:47):
basically took notes that peopletook and Rolodexes that people
had about sales and made themdigital and Epic, which is a
huge company.
With that holds most of the UShealthcare records.
Basically took, well guess what?
Healthcare records that wereheld in, huge organizers and
then batches, and then drawers,and then so on in really, really

(17:08):
large rooms and digitized all ofthat.
And so that was the software andSaaS revolution and what he's
saying that AI will come andoperate on the operations that
are currently done by people.
So it's gonna be booking andrebooking flights, it's gonna be
drafting contracts, it's goingto be negotiating on your
behalf.
It's going to collect payments,by following up with people who

(17:30):
hasn't paid on time, et cetera,et cetera, without any human
intervention.
He gave some very specificexamples, numeric examples, and
he gave an example customerservice.
So he said traditional SaaSmodel charge per seat.
So if you are using Zendesk atabout$115 per seat per month for
the professional suite, you willpay for a large company about

(17:53):
$1.5 million to Zendesk just forthe seats to do your customer
service.
But if you have a thousandperson support team, the cost of
the thousand person support teamis$75 million compared to the
$1.5 million of the software.
However, if Zendesk provides youthe ability to actually provide
customer service, to answerphone calls, to answer chat, to

(18:16):
answer emails, to close tickets,to connect the dots of
everything the clients need,Zendesk can charge you
significantly more, potentiallythree x at$5 million software
spend.
But you will not need the$75million spend on people.
Zendesk will make three x themoney they're making right now,

(18:36):
and the company will be able tolet go of a thousand people
because you will not need themto do customer service anymore.
Now, in many cases, it's notgonna be a all or nothing kind
of game where Ramo highlights.
As an example, nurses in the USare earning$650 billion
annually.
That's more than twice theentire global SaaS market.

(18:58):
Now, can AI replace nurses?
I don't see that likely in thenear future.
However, it can replace some ofwhat nurses do, such as
follow-ups or prescription orpost-surgery pain checks and
things that right now areconsuming a lot of time of what
nurses do and they will not needto do in the future, which means
we will need less nurses than inthe other scenario.

(19:21):
So it's not an all or nothinggame, but it will still require
us to have less nurses, or atleast in the short term, it will
allow us to overcome the needfor additional nurses.
Ramble also played an audio thatis showing a negotiation between
two different voices that arenegotiating and closing a deal
on shipping costs for ashipping.

(19:42):
And it's practically impossibleto tell which one of these two
voices is an AI versus an actualperson, but it got to a final
deal, meaning we're at the pointthat AI voice and AI
understanding of the situationis good enough to have a voice
negotiation between suppliersand people who are buying these
supplies, whether goods orservices, and both sides in

(20:03):
theory could be AI and there'sno reason why they wouldn't be.
Now beyond the saving in costsof ai, ramble stresses that
there are many other advantagesin using ai.
So a few examples that I gave.
One is doing demoralizing roles.
Think about doing collectioncalls.
How many times you gettingreally angry responses, and that

(20:24):
becomes your day-to-day.
That becomes your life.
People just being angry at youand shouting at you and hanging
up the phone on you and callingyou different names.
And now AI will do this, and theAI doesn't care.
It will just keep on doing thisagain and again and again until
they get a response and untilyou actually talk to them and
hopefully pay what you owe.
He also talked about regulatorycompliance.

(20:44):
And at least in theory, the AIwill always be compliant.
It will never go out of script.
It will never do things thatit's not supposed to do.
It will never round corners, andthen you won't get in trouble.
As a company.
He was talking aboutmultilingual support and the 24
7 aspect, 365 of AI that humanscannot just not do.
He was talking aboutseasonality.
So how do you scale a lot ofyour staff around Black Friday

(21:07):
and the holidays, but then youlet them go and then you have to
retrain them again in Septembernext year to get ready for the
holidays again, all of that goesaway when you're using ai and
hence why there's a very, verystrong incentive to use ai, and
that's why there's a very, verystrong incentive for the
companies to develop it assolutions to replace labor
versus just to replace software.

(21:28):
Now you wanna take it moreextreme?
I'm going to remind you acompany that we talked about a
few months back, and thatcompany is called Mechanize and
the goal of Mechanize.
And they raised a lot of moneyfrom a lot of very interesting
people.
They claim that they're going toautomate the entire workforce.
That's their goal.
They're not hiding it.
And they just released a paper.

(21:49):
And the paper is called TheFuture of AI is already written.
And the blog post that theyreleased provide a provocative
analysis of the future asmechanized it is seeing it, and
they're basically saying that wedon't have an option.
They are claiming that the fullautomation of jobs is coming
regardless if they do it or not,just based on everything that

(22:10):
happened in history before.
And they made several veryrelevant historical comparisons.
The first one that theymentioned is that every time
there's been a useful technologyin the past, it was adopted
regardless of what peoplethought about it or felt about
it.
The other thing that theymentioned is that in many cases
in history, there's beenparallel discoveries of new

(22:34):
science, new tools, newtechnologies and so on in
different independent placesaround the world.
So before we had internet thateverybody knew what everybody
else was doing.
It was still happening andthey're claiming it was
happening because once somethingbecame feasible, it was
discovered and then adopted indifferent societies and
civilizations around the world.

(22:55):
And they gave several differentexamples of that.
One of them is that the Aztecsin South America discovered very
similar ways to the Spanish inEurope to do irrigation and use
currency.
That's before the Spanish cameover and found them.
So very similar solutions in twodifferent places around the
world.
And there are, by the way, manyother examples of that.

(23:19):
another example is obviouslynuclear weapons that were
developed and being put to usein different places around the
world, despite the risks thatcomes with it.
And yet we know today that many,many countries in the world have
nuclear weapons and definitelynuclear power.
So what is the claim?
The claim is that since everytime in history something became

(23:42):
helpful and useful, and you candefine what that is, the world
started using it at a largescale regardless of what people
felt about it and regardless howmuch it questions the status quo
that equi that existed before.
And what Mechanize is saying isthat the full automation of the
workforce by AI tools and thenrobots and so on is basically

(24:05):
inevitable.
So they might as well be thefirst one to do it and benefit
from it.
And hence, if you're looking fora job as a software engineer, go
and work for them.
That's the bottom line of thispaper.
So what do I think about all ofthis?
What I think about all of thisis that I sadly think that they
are right in the current way ofthings.

(24:26):
And what do I mean?
In the current way of things.
We are now taking for granteddemocracy and we're taking for
granted capitalism.
Both of these things did notexist, not that many years back
in human history.
Now, this sounds like acatastrophic change of the world
as we know it, but it doesn'thave to be.

(24:47):
So if you think about the lifethat everybody knew in the
Renaissance, everybody knew inthe Middle Ages, everybody knew
in the stone age and so on.
All they knew is that way oflife.
And the change always lookreally scary.
And yet, in each and every oneof these cases, what came
afterwards was significantlybetter.
The vast majority of peopletoday in the world have

(25:07):
significantly better life thanmore or less the entire
population of the world 200years ago, unless you were
extremely wealthy.
And that is just 200 years.
You go back beyond that and youget my point.
Now, do I think this has tofollow the same trend?
No.
There are huge risks and wedon't have a clue what will
follow, and I think UBI isbullshit, but will we find a

(25:31):
solution that might enable us tolive better, more fulfilling
life than we have today withouthaving to work for a living?
I don't know.
I would like to stay optimisticbased on history, but I do see
very, very significant risk,especially in the shorter term
because in the shorter term, wewill not find a solution.

(25:53):
This technology and itsadoption, even though the
adoption is significantly slowerthan the technology itself will
put us in a situation well, wewill have to face realities that
we're not used to facing.
There's gonna be a lot ofunemployment and there's gonna
be a lot of turmoil from asocial perspective, and this
will lead to something differentin the long term.
But in the short term, I thinkwe're heading into some really

(26:15):
turbulent times.
And from that to the secondtopic of the recent releases by
OpenAI and how they're beingperceived by others.
And we'll start with theannouncement from this week, and
I'm quoting Sam Altman's post onX or tweet or whatever it is
called right now.
Now that it's X and on Twitter.

(26:37):
We made ChatGPT prettyrestrictive to make sure we were
being careful with mental healthissues, and that obviously
relates to the suicide that theywere sued about and other
situations that they were reallyharming people that were using
ChatGPT and got the wrongadvice.
So this is me now continuing tothe quote, we realized this made

(26:58):
it less useful slash enjoyableto many users who had no mental
health problems, but given theseriousness of the issue, we
wanted to get this right.
So this goes back to much of theterm oil and the negative
feedback that became very, veryloud as they released GPT five
and took away GPT-4 oh where alot of people say they lost

(27:19):
their best friend and theiradvisor and their personal
relationship with AI wascompletely destroyed because GPT
five doesn't have a personality,but a lot of it was done on
purpose.
So now back to the tweet.
Now that we have been able tomitigate the serious mental
health issues and have newtools, we are going to be able
to safely relax the restrictionsin most cases.

(27:41):
In a few weeks, we plan to putout a new version of ChatGPT
that allows people to have apersonality that behaves more
like what they liked about fouroh.
In parenthesis, we hope it willbe better.
Close parentheses.
If you want your ChatGPT torespond in a very human-like
way, or use a ton of emojis oract like a friend, ChatGPT

(28:05):
should do it in parentheses, butonly if you want it not because
we are usage maxing.
In December as we roll out edgegating more fully and as part of
our treat adult user, likeadults principle, we will allow
even more like erotica forverified adults.
Okay, First of all, let's talkabout what happened on X and on

(28:27):
Reddit as a response to thisannouncement.
And a lot of people went very,very negative, somewhat nuclear
responses, uh, to thisannouncement.
The general consensus was thatOpenAI has said in the beginning
and then later on now thatthey're trying to shift into a
for-profit organization, thattheir main goal is to making the

(28:49):
world a better place with ai.
And yet now they are giving usanother social media platform,
one that might be even worsethan the existing social media
platforms.
And they're talking about therelease from last week of Sora
two as a social app and a pornchat.
One user on X said, and I'mquoting, I don't remember

(29:09):
PornHub, claiming they're goingto make the world a better
place.
Another user shared a videointerview of Sam Altman from 10
weeks ago, in which he said he'sreally proud of OpenAI for not
falling into the trap of tryingto captures people time and
attention on ChatGPT, butactually trying to provide as
much value as possible and whenwas pushed by the interviewer of

(29:32):
an example of what does thatmean?
He literally said that he's veryproud that they still did not
create a sex chat bot on top ofchat GPT.
And here we are 10 weeks laterwhen they are building an app
that will capture as much ofyour time and releasing a sex
chatbot on top of chat GPT.
Now all of that is true, but Iwanna give you my personal

(29:54):
opinion.
I shared in length last weekwhat I think about the social
media app that they releasedwith SOA two.
It.
It scares the hell out of me andit scares the hell out of me
because it is the first timethat a social media platform can
generate the actual content.
If you think about it, yes,humans are asking Sora to
generate it, but Sora cangenerate it on its own.

(30:14):
It doesn't need you to tell it.
All it has to do is to followthe behavior, and as I mentioned
last week, potentially with thecamera, seeing your face and
your expression and what you'reabout to do, and then serve you
a new piece of content thatwould be perfectly aligned with
your personal needs, wants andwishes at that exact second.
Now, is the technology thereright now to do this now?

(30:35):
Because it takes a little longerto render the videos, but this
is what's coming and that isreally, really, really scary.
Like personalized, fabricatedcontent to the wishes and needs
of the individuals, completelyignoring everything else, just
to keep you on the platform aslong as possible.
This is a glimpse of where thisis going.

(30:55):
But the flip side is on theadult chat question.
I have no problem with thataspect as long as they can
really verify that it is adults.
Now, we have had porn sites forlike since the beginning of the
internet, and it doesn't matterwhether I agree with the fact
that we have porn sites or notin modern society.

(31:16):
This is the reality and they arethere and they're available.
And in most cases, any13-year-old kid that has a phone
or 9-year-old kid or girl thathas a phone can access them
because there's no realverification of age when you go
to most of these websites, bythe way, a friend told me that,
I don't really know it.
and the other aspect of this isI have a very serious issue with

(31:39):
open ai, anthropic, Google, orany other large language model
lab being the moral guardians ofthe world.
Why should they decide what isacceptable and not acceptable,
especially across the entireworld, with different societies,
with different beliefs, withdifferent moral codes and so on?
Why should they decide what isright and what is wrong?

(32:00):
So I have no problem with themallowing people to do whatever
they want within what is legal,not what they think is moral
within what is legal, becausethis is how modern society
works, and we treat adults asadults and we allow adults to
make adult decisions, and weallowed it during the internet
era and there's absolutely noreason we will stop doing it in

(32:23):
the AI era.
And I would love to hear youropinion about this.
I would love to hear theopposite side of this
conversation.
So find me on LinkedIn again,you know my name.
I'm Isar Mateis.
I'm on LinkedIn every singleday.
And tell me what you think aboutit.
I'm obviously going to postabout it as well, so we'll have
the opportunity to respond tothat.
This wasn't the only issue withOpenAI in the last couple of

(32:46):
weeks.
They, as we mentioned, releasedSoro two.
SORO two has an opt-out optionfrom an IP perspective.
So you can create videos of anyperson, any known actor, any
character from cartoons aroundthe world, et cetera, et cetera,
et cetera.
And if the companies or thepeople want not to be included,

(33:06):
they need to send a formalrequest to OpenAI saying that
they wanna opt out of theirlikelihood and IP being
available on Soar.
Two, that obviously raised hellfrom many different directions,
especially the Hollywoodcreators.
So the Motion PictureAssociation agencies like
W-M-E-C-A-A-U-T-A, whichrepresenting many, many

(33:28):
different big stars and the SAGaftra.
Association as well all has beengoing very loudly against this
approach by OpenAI, and theirclaim is that OpenAI do not own
the rights to any of that IP andhence should stop this
immediately and replace it to anopt in option where companies

(33:50):
will be able to, if they want toadd their IP into the Sora
universe, but that the way it isright now is a clear
infringement of existingcopyright law.
Sam Atman said a few things.
One of it is that there aredifferent organizations that are
complaining to open ai, thattheir IP is not being shown

(34:11):
enough, basically seeing it aspromotion.
Is that true or not?
We will never know, but that wasthe claim that Sam Altman made.
But the other one, he said thatthey're going to be releasing an
update to Sora two, offering therights holders of this ip.
And I'm quoting more granularcontrol over character
generation, whatever that means.
But this is coming potentiallyaround the corner.

(34:32):
As a reminder, Disney, universaland Warner have sued Midjourney
for very similar issues.
And when they did, I told youit's a very smart move because
they have a much better chanceof beating Midjourney in court
versus OpenAI because OpenAI hassignificantly deeper pockets to
fight them.
And if they win againstMidjourney, then now you have a

(34:54):
precedent that you can use inthe trial against OpenAI or VO
three, et cetera, et cetera now,how will this resolve itself?
I think this will make its wayto the Supreme Court, uh,
because I don't think any of thesides will stop before that.
And I'm not sure that the MotionPicture Association or any of

(35:14):
the other groups, and they'reprobably unified together to go
at this against OpenAI and theother labs, I don't think they
will win.
And the reason I don't thinkthey will win is because this
has been happening forever.
Meaning the user that uses thetool is the one responsible for
whatever they are creating.
If I am now going to use Canvaand I'm gonna copy pictures of

(35:37):
the internet of known charactersand so on, I'm gonna create an
ad and I'm gonna use it onsocial media without getting the
copy rights I need in order tocreate that ad.
And let's say I've used Disneycharacters.
Are Disney going to sue me orthey're going to sue Canva?
They're going to sue me.
They're not going to sue Canva.
Now you can take it further backin time.
When people were creating imageswith paint and pencils, when I

(36:02):
create an image and I sell it,and let's say I'm making a lot
of money from it, and it's animage of Mickey Mouse again,
sorry for sticking with theDisney example.
I'm from Florida.
That makes it very easy for meto think about.
But let's say I created a reallybeautiful painting of Mickey
Mouse and I'm selling it willDisney go after me or will they
go after the people who createdthe paints or hobby lobby for

(36:23):
selling me the paint?
No, they're going to go afterme, the person who is using the
tool in order to infringe ontheir copyright.
I have a very strong feelingthat this is exactly the card
that OpenAI and the other labsare going to play.
And to be completely honest, Ithink it's the right
interpretation of the situation.
The fact that I have a tool thatallows me to break copyright law

(36:46):
doesn't mean I have to use it tobreak copyright law.
And again, this is nothing new.
I could have done this in Canva.
I could have created the videobefore just with a lot more
work.
I could have created imagesbefore, just with a lot more
work.
A little more work.
Again, if I'm using Canva andjust existing pictures from
different places around theinternet, and so this is my

(37:07):
concept.
Now, my view of this, and staywith me for a minute.
I promise it'll make sense.
I think this is a cross betweenSpotify and Web 2.0.
So let me explain what I mean.
If you think about Spotify as weknow it today, as the place we
consume music, and it doesn'thave to be Spotify.
This could be Apple Music orGoogle Music or Alexa or
wherever it is that you consumemusic through streaming.

(37:29):
This is the current state ofthings.
Everybody is more or less happywith it.
So the creators have a hugedistribution channel they never
had before, and they're makingsense on every time their songs
are being played.
But it's played on a very wideaudience and people get to know
them and because they get toknow them, they buy tickets to
their concerts and so on and soforth.
Is that very, very differentthan what it was before?

(37:51):
Yes.
If you remember, we used to buyalbums and CDs of these
creators, and the firstgeneration of distributing music
digitally was the scariest thingthat ever happened to them.
So those of you who are oldenough like me, remember the
peer-to-peer services, the P twoP services that allowed to
exchange music online and betorrent files and servers where

(38:12):
you can go and get literally anysong you wanted, uh, without
paying anybody anything, just bydownloading it.
And then obviously the biggestname that many of you probably
know is.
Napster, right?
So Napster took that conceptthat was kinda like an
underground thing and make itinto a company where you can go
into their website and get anymusic that you want completely
pirated.

(38:32):
That took a while, but it allwent away, and we found a
business model and a usage waythat satisfied both sides, I
think.
Now what does that have to dowith Web 2.0?
Well, web 2.0 was the first timethat allowed anybody to create
content on the web.
Again, those of you who are oldenough who lived in the early
days of the internet, the entirecontent on the internet was

(38:56):
created by really largecompanies, and there was no
social media and you couldn'tpost anything.
You just consumed content of theinternet.
Again, for those of you who areyounger, that doesn't make any
sense, but that was the firstdecade of the internet.
no individual created contentand posted it online.
But now it is the norm.
So what I hope that will happen,and to be fair, I think that

(39:17):
will happen is that the ownersof the IP and the creators of
the platforms will find a way tosatisfy everyone's needs.
In other words, I will be ableto create with AI anything that
I want, and I will pay a littlebit of money for the AI
platform, and some of that moneywill go to pay the owners of the

(39:40):
IP of the things that I amcreating because the platform
will be able to track it anddistribute the revenue
accordingly.
This means that there's gonna bemore Disney characters out there
and more songs by specificmusicians right now, and more
pictures by known artists andmore so on and so forth.
And they will make money off ofit.

(40:02):
They will become more famous andpeople will then consume more of
their stuff.
Now, does that put at risk theoverall mechanism because then
people may consume only this.
Maybe.
do I think there is a way aroundthis?
No, I don't think there's a wayaround this, and the reason I
don't think there's a way aroundthis is because of the open
source universe.
Even if we block Sora two, theopen source tools will allow to

(40:25):
do the same exact thing.
And even if you block thisvariation of that variation of
the open source tool, anyindividual can download the open
source platform to their owncomputer, remove the limitations
of IP creation, and then use itas they wish.
And so I don't think there's away to block it.
And I will think the IP holderswill have to work together in

(40:47):
this new universe, just likethey did with Napster and with
BitTorrent servers.
And we will get to the Spotifyera of creation of everything.
So instead of just consumption,anybody will be able to create
anything.
If and if there is IP involvedin it, you may need to pay a
little more to create it andpart of that money will go to

(41:08):
the owners of the ip.
Now, will this happen?
I don't have a clue.
Does it make sense to me thatwill happen?
Yes.
Do I want it?
Do I wish it will happen?
I do.
I don't know about the bigowners of the IP right now, but
again, as I said, I don't thinkthey have much of a choice.
Now, staying on this topic ofSora two, Japan's government

(41:28):
just issued a call for open AIto, and I'm quoting refrain from
corporate infringement in Sorotwo related specifically to
anime characters.
Which they defined as, and I'mquoting again, irreplaceable
treasures, that Japan boast tothe world And they're

(41:48):
emphasizing what they're callingthe irreplaceable role of anime
to global culture and economy.
So where do we stand right nowon OpenAI?
We stand in a situation that intwo weeks they literally went
from, we are all for the futuregood of humanity.
We are not going to do thingsthat will just try to keep you

(42:09):
on the app as long as possibleto, as I mentioned, giving us a
new really scary social app andpotentially the ability to do
porn.
Chats on chat.
GPTI already told you what Ithink about it.
But since we mentioned, agovernment and the CHI and the
Japanese government approachingopen ai, I will mention one more

(42:30):
thing on this segment that isnot completely related, but it
is tied to regulations and soon, and that is california just
passed a call called Senate Billtwo four three, and it was
signed into law on October 13th.
And what this bill does is itrequires chatbots to clearly
notify users that they are AIand not human, and mandate

(42:50):
annual reports on the safeguardsin the situation that they have
on negative things such assuicidal ideation.
The idea here is obviouslytwofold.
One is to make it very, veryclear to users that they're
talking to an AI versus a personin order to reduce the potential
dependency or trustworthiness ofthese tools.

(43:12):
And the other is to generate away to monitor what's actually
happening in those platforms,especially when it's life
threatening.
Meaning suicide preventionefforts.
Now do I agree with that?
And do I think it will play asignificant role in the future?
So yes, I obviously agree withthat, right?
It puts reasonable guardrailswithout really stopping the
technology from moving forward.

(43:34):
and I think it makes perfectsense.
Do I really think it will play asignificant role in.
Doing what it is trying to do.
I think time will tell.
I think telling people it's anAI versus a human doesn't
matter.
I think people who chat on thesechatting platforms know it's ai.
They don't think it's a human.
And even if you keep onreminding it to them, it doesn't

(43:55):
matter because if you developemotions as an outcome of a
conversation, the fact thatsomebody will tell you that your
emotions came from somethingthat's not human will not make a
difference.
We get emotional when we watch aDisney cartoon movie.
Just think about how you feltand when you saw Bambi for the
first time and the firehappening, uh, and what happened

(44:16):
to his parents, right?
We shed a tear and we becomevery, very emotional and we know
it's not real.
It's a cartoon on a screen, andyet it drives an emotion.
If that emotion continues for along period of time, you will
develop some kind of anemotional attachment to that
thing, regardless of whether itis human or not, despite the
fact that, you know, it's nothuman.

(44:38):
So I think it's not going toprevent that.
I think it has no chance toprevent that just because the
way we are wired as far asshowing people what is the
impact of that, and forcingcompanies to share publicly what
is the overall situation from asocial impact perspective,
especially at its extremes.
I think that is very, veryimportant because it will force

(44:58):
companies to take the rightmeasures to prevent these things
from happening.
And now to the third deep diveof this week.
And that's what happened inAgent Force this week.
Agent Force is what used to beDreamforce, which is Salesforce
largest conference of the year,announcing all their biggest
announcement.
Again, they've changed the namelast year because they wanted to
show the world that they're allin on AI and focusing on

(45:21):
developing agents.
Now, why is this very importantand why is it a deep dive topic?
Because it connects directly tothe first segment that we had
and how will AI change softwareand or how we work.
And so in the first agent force,mark Benioff has been very loud
and vocal about how fast thisshift is going to happen, how

(45:42):
quickly it will take over theworld, and how generic AI tools
like ChatGPT are bullshit andhas no value in the real world.
And this year he acknowledgedand had a complete shift in his
rhetoric.
He is now stating that utilizingAI effectively and implementing
it on a enterprise level takestime for companies and it's not

(46:06):
a quick flick of a button andthen you're up and running.
He, he also agrees with what is,I think, a consensus that AI
technology is definitelyoutpacing consumer adoption.
And as I mentioned, heacknowledged to an extent the
real competition, and that waskinda like the underlying clear

(46:28):
tone, that it is very obviousthat Salesforce sees the moves
by OpenAI and Anthropic as athreat to their business, as
they're growing their influence,and as they're integrating into
more and more tools, and it'snow more or less, if you can't
beat them, join them kind ofapproach, versus this is
bullshit.

(46:48):
It will never be relevant.
You have to stick with what weare giving you.
Now Salesforce shares are down26% year to date, and a lot more
than that since its peak onJanuary 28th, which is
definitely not a good sign asfar as what people thinking
about the chances of Salesforce,figuring out how to deliver
value and how to grow sales inthe current situation of things.

(47:11):
Now, what Mark Benioff says isthat he suggests that Salesforce
competitive advantage lies inits deeper customer relationship
and the fact that they hold, andI'm quoting fundamental backbone
of their mission criticaloperations end quote, which led
them to announce their newversion of the product that they
called Agent 4 360.

(47:33):
So what is Agent 4 360?
It is basically their way tointegrate everything,
Salesforce, all the differentcomponents under one umbrella
from a data perspective andallowing the users to build
agents that can connect and talkto all these tools at the same
time, or as they're calling it,it's the connective tissue that

(47:54):
brings together sales,marketing, commerce, slack,
Tableau, MuleSoft, andeverything else Salesforce under
one agentic layer.
This is obviously an extremelypowerful promise, if it is
actually doable.
So let's talk, first of all,what is the goal of Agent Force
360?
And then talk about what isactually included in it.

(48:14):
What they are creating is anagentic enterprise solution end
to end with all the data and allthe different tools that they
have, which is a very powerfulpromise that I think every
enterprise that is usingSalesforce or is that, is using
any other set of tools wouldlike to have Now, is it
complete?
No, it's not complete becausethere are many tools that are
not Salesforce, thatorganizations are using.

(48:37):
You still have an ERP, you stillhave, uh, outlook or Gmail to
send your emails, and you stillhave databases for other
platforms, and you have.
Other tools, but at least on thestuff that is in within the
Salesforce universe, it issupports to unify all of it.
So what comes with it?
it comes many different tools.
One of them is called AgentScript, which is allow you to

(48:57):
prompt and create scripts thatallows users to create agents.
There are many different ways tocreate agents.
I'm not a hundred percent sureon the differences yet.
underlying underneath all ofthis are the reasoning models
from anthropic Open AI andGemini that can power your
agents to be able to be verypowerful and capable because of
their thinking capabilities.

(49:18):
They're also created agents,force Builder, which is another
new tool, and most of thesetools are gonna be released to
beta in November, so they're notreleased yet.
And this tool is allowing usersto build, test, and deploy
agents from a singular spot.
They also are releasing agentforce vibes, which allows you to
vibe, code, differentapplications that talk to all

(49:40):
these underlying tools,capabilities, and agents.
They're also releasing Data 360,which is a new variation of Data
Cloud, which is a intelligencelayer for their data platform
that allows agents to haveaccess to clean and reliable
data and context for everythingthat you're building.
And they're also releasing morethan 300 industry ready agents.

(50:03):
So stuff you don't have todevelop, you can just take them
off the shelf.
And the goal is to make themplug and play into different
industries like healthcare,financial services, and
manufacturing.
Now, when I mentioned, if youcan't beat them, join them.
If you go back to last year'srhetoric about open AI and
ChatGPT Versus what theyannounced this year.
So this year you versus whatthey announced this year, you

(50:25):
see that there's a very, very,very big difference.
And what they announced is thatusers of OpenAI ChatGPT, that
has Salesforce licenses, can nowconnect their chat UPT and give
it full access to everything onthe Salesforce app data,
including CRM records.
And it can control Agent Forcetools straight from chat GPT or

(50:47):
as I said, chat, GPT will beable to, and I'm quoting, tap
into powerful enterprise gradeai, right at their preferred
surface environment.
Basically, you'll be able to usechat GPT to control and
understand and connect witheverything on Salesforce.
Now they have positioned Slackas the home for all these
agents.

(51:07):
So basically Slack becomes yourcommunication world with the
agents behind the scenes.
And so any communication thatyou want to do with any data
will be done just like you'redoing it today.
But instead of to talking tohumans, you will talk to agents
behind the scenes.
I think this makes perfectsense.
Instead of trying to introduce awhole new way of communication,
let people communicate asthey're communicating today,

(51:29):
just provide them significantlymore value through an existing
platform.
And I think that's a very smartmove by OpenAI.
I already have several differentintegrations of AI into Slack
and not natively, but just stuffthat I created, uh, including
N8N and Make and stuff likethat, that talk into, uh,
specific channels that I have inSlack.
And it's very useful andintuitive and making it out of

(51:50):
the box makes a lot of sense.
Now, the current statistics isnot very promising and it's not
very positive, which might beone of the reasons why their
stock keeps on coming down thisyear.
But they currently have 12,000customers that are using or
experimenting with Agent Force,so we don't really know how many
of them are in a mature phase ofactually using it.

(52:11):
Now that sounds like a prettylarge number.
But the reality is they haveabout 150,000 customers.
So after over a year of tryingto deploy this and getting into
less than 10%, where many ofthese, and again, we don't know
how many are just experimentingand not actually deploying Agent
Force, uh, that's not a verypositive sign as far as adoption

(52:33):
of this new technology.
But again, agent Force 360 is awhole different beast and
animal, both in means of itsability to connect to the data
as well as in the organization'sability to easily create and
deploy agents.
So this might change it movingforward.
Now to show how amazing agentforce is, they gave multiple
examples of large organizationswho are already deploying and

(52:58):
getting benefits and immediateROI by deploying Agent Force,
these included Dell and FedExand William Sonoma and Pandora
and PepsiCo and others.
So they are showing specificcases in which large
organizations have deployed itand are seeing real ROI.
We're obviously going to seemore and more example like this

(53:19):
moving forward, but as of rightnow, I think Salesforce are
making the right moves toprovide value.
And if we connect it to the veryfirst segment that we had in
which we shared that in theorySalesforce will be able to be
your sales team and yourmarketing team, then if they
move in that direction, and ifthey do that effectively, they

(53:39):
can not just save Salesforce,but make it one of the most
successful companies in history.
Tam will tell if they are goingin that direction and how
effective will they be inactually deploying it in a
successful way.
But they definitely have theopportunity because they own the
data and they're connected toeverything across many, many
different companies around theworld.

(53:59):
Those of you who have taken mycourse, you know I have my five
rules for success in the AI era,and one of them is that there
are two ways to win in the AIera.
One of them is having access andowning the proprietary data.
And in this particular case, itis very clear that Salesforce
owns the proprietary data of ahuge amount of companies around
the world, which gives them inthe perfect position in order to

(54:23):
leverage that data to allowthese companies to be more
effective with that data, whichmay or may not come at the price
of the employees of thecompanies.
Now two hour rapid fire items.
And I'm going to start withApple that took another hit this
week.
Kay Young, who just recentlywere promoted to head of
Apple's, a KI team has leftApple for meta after just a few

(54:47):
weeks in the new role.
So the a KI team stands foranswers, knowledge, and
Information Group that wastasked with enabling Siri to
fetch real-time web data andconversational queries.
Basically making a smarter AIbased Siri.
So the person who is just madein charge, just left for meta.
This is just another nail in thecoffin of apple's AI

(55:10):
initiatives.
It literally just doesn't lookgood again and again and again.
And this will eventually come tobite them and will start
impacting the stock price ofApple.
Because yes, it's been going upso far.
I don't really understand why.
Uh, yes, they can release newdevices and devices are okay.
I guess they're better than theprevious models, but this is not

(55:31):
the new game.
And in the new game, they havelost time and time again.
I'm really surprised theyhaven't bought one of the
leading labs, or at leastperplexity, which was in
conversation with'em so far inorder to solve this problem.
I don't think they will have achoice.
The other choice would be to saywe are not in the AI game at
all.
Basically just announce it andsaying, we are going to keep on

(55:53):
building some of the besthardware on the planet and we
will integrate other people's AIinto it in the most effective
way.
I think that will take a shortterm hit in their stock.
But I think in the long run,that may be another approach to
basically say, we're not eventrying to compete on this.
We will just keep on buildinghardware.
That's what we know how to do.
or like I said, go and buyphilanthropic, which is gonna be
very, very costly, but doable byApple.

(56:16):
But that's not the only personwho has left a known company to
join Meta this week, AndrewUllo, who is one of the
co-founders of Thinking MachineLabs, the company that was
founded by Mira Moira fromOpenAI, who raised a crazy
amount of money before evenannouncing exactly what product
they're going to release.

(56:36):
Well, he's leaving for Meira aswell for as a spokesperson for
the company.
Said personal reasons.
No, this is not the firstattempt of meta to bring him
over to the meta team.
The first attempt was in Augustwith a crazy offer that
presumably was about$1.5 billionin stock and incentives

(56:59):
performance over time.
Now, meta said that's inaccurateand ridiculous at the time that
this, was, that the article inthe Wall Street Journal shared
that amount.
But at that time tulo refusedthat offer whatever the offer
was.
Uh, it obviously was verysignificant and now he is moving
over.
Uh, so those personal reasonsmight be I'm going to become a

(57:22):
billionaire overnight.
So that's a good enough personalreason to change, from one
company to the other, especiallywhen the other company already
has a product and distributionand compute a lot more than your
current company has.
Now, a little bit of history whyTulo is so important.
He spent 11 years at meta'sFacebook AI research, and then

(57:42):
at Open ai, and then obviouslynow at Thinking Machines.
So he brings a real seriousdepth and experience and
knowledge about the META'SFacebook universe back into the
team.
In a very interesting piece ofnews for this week, OpenAI and
Broadcom announced a multi-year.

(58:02):
Partnership to generate 10gigawatts of custom AI
accelerators for open ai.
So this is now custom chips.
This is not open AI buyingstandard off the shelf chips
from Nvidia.
These are going to be builtbased on their blueprint and the
deployment is supposed to startat late 2026 and continue

(58:22):
through 2029 with newinvestments that are coming from
unshared sources.
At this point, this is gonna bethe first time that OpenAI will
have its own processor aftertrying to do this and announcing
of trying to do this severaltimes in the past.
Now this is a really broadpartnership with Arm Design
specialized CPUs to pair withthese Broadcom chips.

(58:45):
TSMC is going to fabricate thechips on their latest
technology, and SoftBank willobviously play a role in this,
in financing the whole thing.
So a very large partnership thatis somewhat built to hedge
against the complete dependencyon Nvidia right now in future
growth.
And we have seen other similarpartnerships in the past few

(59:08):
weeks announced by OpenAI withother tools.
But this is the first time thisis gonna be a chip that's going
to be developed to the blueprintthat OpenAI will put together.
But OpenAI are not the onlycompany who made similar
announcement this week.
Meta is partnering with Designcompany ARM to create AI
capabilities and AI chips toMeta's needs.

(59:29):
This obviously aligns with thecrazy development right now of
multiple new data centers.
Meta themselves has severaldifferent large projects.
The two biggest ones arePrometheus, which is a huge
multi gigawatt data center thatis supposed to go online in 2027
in Ohio and Hyperion, which is afive gigawatts that is supposed

(59:50):
to go live in northwestLouisiana in completion by 2030.
And so having their own designchips will obviously allow them
to do things their way andreduce cost and reduce
dependency on other companies.
I think we're gonna start seeingmore of that across the board
with companies with the largelabs developing their own chips

(01:00:10):
in order to have less dependencyon Nvidia and third party
providers.
There was a very interestingarticle this week on the
Information, which is a onlinemagazine that I really like.
They bring a lot of good foodfor thought and information from
different sources, includingabout ai, and they're sharing
about the current huge dilemmathat Amazon has right now.

(01:00:32):
I shared with you last week thatWalmart signed a multi-year
agreement with OpenAI, andthey're becoming the first
really large company to allowcheckout of any Walmart product
on the ChatGPT platform.
Basically, you can go and searchany product that Walmart has on
ChatGPT and be able to buy itand check out right there and
then without ever visiting theWalmart website or app.

(01:00:55):
This took Walmart's stock upalmost 5%, which in Walmart's
value is a crazy amount ofmoney.
At the day of the announcementand at the same time, Amazon
stock fell 1% and if you look ata longer period of time,
Target's stock went down 34%year to date.
So where does that put Amazon?

(01:01:15):
Amazon today is definitely thekingpin of e-commerce in the US
They're holding about 40% marketshare of us E-commerce.
That's a crazy number.
And now they need to make achoice.
They either do the same thingthat Walmart did, and allow

(01:01:36):
people to shop on Amazon, righton ChatGPT or allow 800 million
weekly users on chat g PT rightnow, which is growing every
week, so that it's gonna be abillion plus probably as we get
into 2026, allow them to onlyshop things from Walmart and not
have access to Amazon, whichwill definitely eat into their

(01:01:59):
market share.
The biggest problem I thinkAmazon has with this is that a
big part of Amazon revenuedoesn't come from the revenue
sharing they have on everyspecific product, but comes from
ads on the platform.
People who want to promote theirproducts on Amazon are paying a
lot of money for placement, andif they go down the path of

(01:02:20):
allowing people to shop Amazonproducts on ChatGPT, and
obviously later on otherplatforms as well as agents,
they will lose a part or maybeall of that revenue over time,
which is definitely not goodfrom a stock price perspective.
Now, how will this turn out forAmazon?
I don't know.
It will be very interesting tosee how they play this.

(01:02:43):
I will say something aboutAmazon's previous attempt at
this.
So if you think about it, mostpeople before Amazon were
shopping and still many peopledo shopping for goods on Google.
And then from Google going tothe different sites and Amazon,
were able to change thatparadigm in many, many, many

(01:03:04):
different people in the world,myself included.
And definitely in the us.
Are going to Amazon to look forstuff.
And they're not looking onGoogle at all.
So it's maybe the only case thatin the Western Hemisphere people
are going to a differentplatform than Google to search
for stuff.
And that is the Amazon example.
Can they do this in the open AIand agent era?

(01:03:26):
Time will tell.
I would argue that in the longrun, they don't stand a chance
because once everybody's used tousing their own agent to do
everything for them, and when Isay everything, I mean
everything digital.
Then they will go to that agentto find stuff to shop, and that
thing will have to find it onthe internet.
And then Amazon will not have achoice but to shift to that
model.

(01:03:47):
But in the short term, it willbe very interesting to see how
they play this.
Now, last week I shared with youthat OpenAI is very clearly
going after world dominationacross literally everything they
can put their hands on.
So this week it's became knownthat they are pitching many
companies to partner with themon sign in with ChatGPT.

(01:04:08):
So similar how we have today,sign in with Google or sign in
with Microsoft or sign in withAmazon.
Uh, with the biggest one.
Obviously being Google, you willbe able to sign into third party
platforms using the chat GPTconnection.
Now connect that with theannouncement from last week with
the announcement of apps in chatinside of ChatGPT, and you
understand where this is going.

(01:04:29):
If by default your sign into theapplication is with ChatGPT and
behind the scenes you canconnect to the app and use it
within the ChatGPT ecosystem,that will evolve more and more
and more.
Then OpenAI has their fingersinto more and more and more
places, and they will become thesolution that people will go to
for basically everything theyneed.

(01:04:50):
There are similar solutions tothat in China, if you think
about it.
They have single large apps thatyou can use for everything,
including payments and loginsand transportation and literally
everything you want.
and it seems like OpenAI aremoving in that direction in a
very aggressive way.
Connecting back to what wetalked about in the beginning.
Zendesk is moving veryaggressively to develop agents

(01:05:13):
on the Zendesk platform and intheir AI summit, they shared
that their new autonomous AIagent is designed to
independently resolve 80% ofcustomer support issues that
they are seeing right now.
So think about it.
They have the data of exactlywhat is happening from a
customer service perspectiveacross well, a huge amount of

(01:05:33):
companies that are using theZenex platform, and now they can
build agents to handle thesesituations, to build it to spec
of exactly what the needs are.
Which will allow them going backto what we discussed in the
beginning to become theworkforce for customer service
versus just the tech SaaSplatform to hold the data for

(01:05:55):
customer service.
Now, in addition to be able tohandle the 80%, they're also
rolling out co-pilots that willallow the remaining 20% of the
people, to collaborate with AIto deliver customer service
resolution faster, better, andcheaper.
They've announced, you know, anadmin layer to control all the
agents, a voice-based agents,and a lot of analytics in the

(01:06:16):
backend to show exactly what ishappening.
So whether we like it or not,this is coming and it is coming
very, very fast.
And even if they're exaggeratingby 100% and it's going to solve
only 40% of use cases across theentire user base of Zendesk, it
is very, very significant.

(01:06:36):
It will obviously improve overtime.
A company called Finiti, spelledwith a capital D, and then
Infiniti after that has justreleased Caffeine, which is
another vibe coding platform.
They're claiming it is slightlydifferent than the existing Vibe
coding platforms.
They have their own developmentlanguage for a while.
So this company is not a newcompany and the goal is to allow

(01:06:59):
people to very quickly build webapplications and deploy them at
a enterprise level of certaintyand security.
whether that is really differentthan what we have right now or
not, I don't know, but that'sanother vibe coding solution.
And we are going to have a bunchof these very successful.
My go-to right now are mostlylovable and rep, but there are

(01:07:22):
many others out there.
And this is just anothercontender this very hot part of
the competition of the AImarket, staying on new releases
grok.
So Elon Musk's company justreleased Grok, imagine version
0.9, is a image and videocreation platform.

(01:07:42):
And the interesting thing aboutthis is that it generates videos
extremely fast from either textor images.
So from a speed perspective, youcan generate video on Grok
faster than any other platformout there.
And the very initial examplesthat I've seen are showing very
solid realism and motion as faras the quality of the video that

(01:08:06):
it generates.
Now, I shared with you in thelast couple of weeks that Grok
four Fast, their smaller, fastermodel is currently the most
efficient model, more or lessacross the board.
So from a speed and cost tovalue perspective, they are
alone in the top right side ofthe quadrant of being able to

(01:08:26):
deliver value fast and cheap.
And this is just another exampleof that.
Only on the image and videocreation side of the universe
will it be able to compete withso two and VO three the next
generation of tools, time willtell, but they're definitely
doing the right things to placethemselves as the most cost
effective fast solution in theAI race as of right now.

(01:08:49):
And speaking of VO three, Googlejust unveiled VO 3.1, which is a
jump forward from the previousmodel that is now really, really
old, and it was released just afew months ago.
So they released two models, VO3.1 and VO 3.1 fast.
Just like with the previousversion of just VO three.
It is available through theGemini API, the Google AI Studio

(01:09:11):
and Vertex ai.
And it's also available on theregular apps of Gemini and Flow.
It provides enhanced audio andvisuals compared to the previous
model.
It allows to use three referenceimage characters to create
character and style consistencyacross video and shot.
It includes the ability toextend scenes so you can take an

(01:09:32):
existing video and extend itover time, up to a minute or
longer by generating new clipsto continue the original clip.
So basically allowing you tocreate longer and longer videos,
which is something that most ofthese tools did not know how to
do.
You had to create the separatecuts and then bring them
together in an external editor,and now you can create much

(01:09:53):
longer sections and segments.
That was not doable before.
Definitely not at that quality.
It creates a smooth transitionbetween a first frame and last
frame of a video and knows howto work with it, including the
audio that has to come with it.
And they have a few new featuresthat did not exist before that
are going back to the toolingside of things with the ability

(01:10:13):
to crop and zoom and do thingsthat were not possible before.
And that goes back to what Isaid many times before.
We are getting to the levelwhere these tools are getting to
the point that they're allreally, really good.
And what's gonna make thedifference is not necessarily
the marginal difference in thequality of the output, in this
case video, but how easy,effective, and useful the
tooling is going to be.

(01:10:34):
Meaning how easy will it be fora user to get the output that
they want versus how good thequality is.
Because the quality is gonna bemore than acceptable across all
these different tools.
And these are the first moves byGoogle to move more in that
direction.
Now the cool thing, if you wannalearn how to use Google VO three
very, very quickly and get verypowerful capabilities, Google

(01:10:57):
also released VO threeprompting, VO 3.1 prompting
guide.
So you can go to the guide, copyand paste example, and use their
five part prompt formula tocreate incredible outputs that
look completely realistic or notrealistic, uh, but basically
look however you want.

(01:11:17):
And there's gonna be a link tothat in the show notes and
obviously in our newsletter aswell.
Now Microsoft is rolling out abig update to the copilot app on
Windows, and the idea is that itwill be able to seamlessly
produce new documents and so onout of thin air instead of just

(01:11:37):
helping you with editing andstuff like that before.
And this includes full Worddocument, Excel, and PowerPoint
presentations, as well as lateron PDF files directly from chats
and prompts, bypassing basicallythe current apps.
So this is released for initialtesting.
It's still not widespread yet,but this is the direction that

(01:11:59):
it is going.
So the days in.
The dream that I thought willcome true sometime in late 2024
might finally happen in 2026,where there's gonna be one
unified solution for Microsoftand one unified solution for
Google versus standalone appsthat are somewhat useful on
very, very specific use cases,but definitely not as useful as

(01:12:21):
they can be if they can accesseverything across the board.
And I am really interested tosee it's finally coming
together.
myself as a Google user.
I'm obviously more interested onthe Gemini side of the unified
platform, but there's definitelya huge audience for a unified
copilot that can get to all yourdata across everything
Microsoft, and create outputsacross the entire Microsoft

(01:12:43):
universe.
Hugging Face did an interestinganalysis on all the different
things that are currentlyavailable on Hugging Face.
They released that on October13th, and it includes the
platform's 50 most downloadedopen source entities that
account for over 80% of all ofthe downloads on the platform.

(01:13:04):
So the 80 20 Rule works right?
It's just 50 entities, but theyaccount for 80% of the downloads
across thousands or tens ofthousands of things.
You can currently download fromhugging face.
So they come from 20 companies,10 universities, 16 individuals,
and they together drove 36.45billion downloads.

(01:13:25):
Now small models are crushing itwith 92.48% of overall models
are less than 1 billionparameters, which is showing you
that while we're all looking andchasing the big models under the
hood, behind the scenes, peopleare building different
applications on open sourcetools are actually preferring
much smaller models becausethey're fast and cheap to run.

(01:13:49):
Now from a modality perspective,computer vision is at 21%, audio
is at 15% multimodality, sotools that can do all of them is
at 3.3%.
I think it's just not asdeveloped on the open source
world yet, but this gives you anidea with the, obviously on top
is text models.
Still English dominates theoverall languages of these

(01:14:10):
different models.
US entities from the number ofentities are leading, but this
month was the first month thatChinese models were downloaded
more than US models.
For the first time ever, therehas been more downloads of open
source Chinese models than USmodels, despite the fact that
the US have many more entitieson the list, way more than

(01:14:31):
everybody else.
What does that tell us?
It gives us an idea of what'sactually happening behind the
scenes when developers want tochoose open source models for
the different things thatthey're developing, showing you
that there's a very live andhealthy open source environment,
and showing you that China isrising very aggressively on that
field on purpose.
All the big releases that theyhave done have been open source,

(01:14:53):
and they have placed themselvesas the leader in open source
model development.
And now for the final and reallyexciting news that I promised
you for the end of this episode,which is Google Gemma powered AI
tool that was custom created forthis purpose has been able to
create a breakthrough in.
Potentially providing a new pathfor cancer treatment.

(01:15:17):
So I am not a scientist and I'mdefinitely not an expert on
cancer, but I'll try to explainto you what I understand from
reading this paper.
So the model is called C two Sdash Scale 27 B model, which is
designed for single cellanalysis.
And what it's trying to do isthe following, apparently there

(01:15:38):
are two types of cancer.
There's cold tumors, and thenthere's tumors who are active
and these cold tumors evadedetection of our immune system
and T cells.
So the idea was to come up withdifferent types of chemical
compounds that will triggerthese tumors in order to being

(01:15:59):
detected by human T cells andhence attacked earlier in the
process and hence potentiallyhelping fight cancer in earlier
stages.
And what this model was able todo is it was able to identify
specific candidates as far ascompounds that can do this,
checking over 4,000 differentoptions, and then recommending

(01:16:21):
the ones that are most likely todo that.
The ones that the AI has foundhave been then tested in an
actual lab with an actual celland was proven to actually work.
This is a true new breakthroughin science that makes a huge
difference on human life.
AI was able to do potentiallyfor the first time ever.

(01:16:45):
Now, is this really gonna helpus cure cancer better or not as
unclear?
This is was one building blockout of many others.
So there's a lot of other piecesin this puzzle to take this from
a concept to an actual solutionthat can be used.
But what is showing that novelscientific discovery can be done
either completely by ai,indefinitely with the assistance

(01:17:05):
of ai, which can then help solvesome of the biggest problems and
the biggest issues that we havein our world today, such as
cancer, other disease, globalwarming, power consumption, et
cetera, et cetera, et cetera.
So kudos to Google for creatingthis model and working with
scientists to do this, and let'sreally hope that together we can

(01:17:26):
create a better future with ai.
Will we be back on Tuesday withanother how to podcast, where
we'll teach you with some of thebest experts in the world how to
do something that you can do inyour business with ai.
If you are finding this podcasthelpful, please hit the
subscribe button so you don'tmiss any episode.
Do it right now.
Pull up your phone and do this,and while you're at it, share it

(01:17:47):
with other people.
There's a share button on yourpodcast player.
Just click on that and add a fewnames of people that you know
can benefit from this.
They will learn a lot.
The world has a better chance ofbeing successful with ai, and I
will really appreciate it.
And until next time we talk,have a great rest of your
weekend.

All Episodes

233 | ChatGPT porn chat is coming, Job quake: 100 million roles at risk this decade + Mechanize’s “total workforce automation”, the biggest lessons from AgentForce, AI cancer-fighting novel breakthrough, and more AI news for Oct 17, 2025

Episode Transcript

Popular Podcasts

Stuff You Should Know

Dateline NBC

On Purpose with Jay Shetty

.css-15opob5{left:0;position:absolute;top:0.8rem;} All Episodes

Episode Transcript

Popular Podcasts

.css-r6mb8g{margin:0;word-break:break-word;display:-webkit-box;-webkit-box-orient:vertical;box-orient:vertical;-webkit-line-clamp:1;overflow:hidden;}Stuff You Should Know

Dateline NBC

On Purpose with Jay Shetty

All Episodes

Stuff You Should Know