Clara's AI Hiring Claims Questioned
Key Points
- The speaker downplays OpenAI’s new search feature, saying it’s a modest improvement rather than a breakthrough innovation.
- Clara is aggressively promoting AI‑driven automation and the elimination of up to 2,000 jobs to impress investors and defend margins as it prepares for an IPO.
- Despite publicly “pausing hiring,” Clara’s job listings still show dozens of senior engineering openings, suggesting a gap between its messaging and reality.
- The company’s push to automate processes like Salesforce is framed as a way to recover its 2021 valuation peak of $45 billion, now reduced to about $14 billion amid broader market compression.
- The speaker also highlights Ilia Suer’s presentation at the NORS conference in Vancouver as another noteworthy, yet under‑reported, AI development.
Full Transcript
# Clara's AI Hiring Claims Questioned **Source:** [https://www.youtube.com/watch?v=iuZyIJ-VzRM](https://www.youtube.com/watch?v=iuZyIJ-VzRM) **Duration:** 00:08:39 ## Summary - The speaker downplays OpenAI’s new search feature, saying it’s a modest improvement rather than a breakthrough innovation. - Clara is aggressively promoting AI‑driven automation and the elimination of up to 2,000 jobs to impress investors and defend margins as it prepares for an IPO. - Despite publicly “pausing hiring,” Clara’s job listings still show dozens of senior engineering openings, suggesting a gap between its messaging and reality. - The company’s push to automate processes like Salesforce is framed as a way to recover its 2021 valuation peak of $45 billion, now reduced to about $14 billion amid broader market compression. - The speaker also highlights Ilia Suer’s presentation at the NORS conference in Vancouver as another noteworthy, yet under‑reported, AI development. ## Sections - [00:00:00](https://www.youtube.com/watch?v=iuZyIJ-VzRM&t=0s) **Clara's AI Hiring & IPO Pressure** - The speaker downplays OpenAI's new search feature and argues that Clara's controversial hiring cuts and ambitious claims of automating thousands of jobs—driven by valuation compression and IPO pressure—constitute a far more significant AI development. ## Full Transcript
two pieces of AI news that I think
matter and one that doesn't the one that
doesn't first I don't think open AI
search is going to go down as a massive
uh Innovation they already search
they're searching in advanced voice mode
that's really the end of it I think it's
helpful but it's not that big a piece of
news but two things I think that are big
that aren't getting the headlines they
should one is the way Clara is handling
their hiring practices and automating AI
conversation I think a poor example of
how to handle a delicate situation and I
think it's driven by Claro's business
performance and their pressure to IPO
and we're going to talk about it briefly
and then I'll get to Ilia suer and his
speech at the nors conference in
Vancouver after that so first Clara
fundamentally Clara is trying to recover
their valuation from 2021 when they were
at $45 billion that is what the lens you
should use for everything they do
they're currently valued at 14 billion
they obviously have a long way to go
value compression is just a reality for
most companies that had a valuation in
2021 part of why Clara is being so
aggressive then with their claims about
automating away Salesforce and now uh
today automating away uh 2,000 jobs I
think that's fine like they're they're
going to make those claims because they
want to impress investors with their
efficiency in a low margin by now pay
later business with a a lot of bad debt
it makes a ton of sense for them
business-wise to argue that they're
extremely efficient because it defends
their margin and it helps them project
defensive margin when they're looking to
IPO that doesn't mean that it's actually
as substantive as they claim that
remains to be seen it is possible that
Clara will end up aing automating away a
significant part of their workflow they
have unlike many SAS providers very
defined business logic they have very
clear product requirements it's actually
one of the lower ambiguity SAS
businesses out there you buy now you pay
later like there's a lot of detail in
bad debt and how you manage bad debt
I've looked at that briefly in another
Walk of Life but overall from a business
perspective it's one of the clearer
playbooks to
run so I could see if they really wanted
to commit to it that it might be one of
the earlier places where you start to
see some labor replacement with AI Etc
but what I don't
appreciate is CLA claiming they are
already doing this when their actual
jobs page doesn't reflect that reality
they are saying they've paused hiring
for example but you can find 54 open
positions on Clara's website right now
and on LinkedIn and they're not they're
not meaningless positions they're hiring
senior software
Engineers so if you're going to claim
that you are pausing
hiring and then publicly not live up to
that promise I think it's right for
people to ask questions about why you're
making such aggressive
statements when the reality doesn't live
up to what you're claiming so I think
that the reason I call this out is
strategic is that in 2025 we are going
to see a lot more statements from
companies that feel like clar's
statement because it's going to be very
fashionable for companies to claim huge
gains from AI efficiencies we need to
ask where the reality is we need to ask
where people are actually delivering
gains with AI in production versus where
they are making a PR statement and right
now it's really difficult to see what
clar is actually delivering and we're
just going to have to wait and see and I
think that the lesson I take away from
this is to be a little bit skeptical
about AI automation claims and jobs
heading into
2025 it's like prove it right moving on
to number two moving on to Ilia and his
speech at norp in
Vancouver so Ilia is the grandfather of
AI not the only one there's a lot of
people who can make that claim and IL is
a young guy so he's he much less gray
than me but he has been instrumental in
building AI he was a co-founder at open
AI he has been uh critical to the
development of large language models
since 2014 so they invited him back a
decade later at NPS to reflect on sort
of how things have come this far what's
up ahead very classic thing right and
now he's a Founder at safe super
intelligence so he is intimately
involved in building a motion to Super
intelligence which is usually defined as
uh an intelligence that is great enough
to run an entire organization's worth of
value through AI without any people
whatsoever so imagine like your startup
of 200 is just an
AI so he comes and mostly what people
report on is his quote that the internet
is like oil it's a non-renewable
resource we've already used it for
pre-training and it's
gone and so people talk about that right
and that's his it's framed as his
comment on the great uh pre-training
controversy where we are wondering if
we're hitting a pre- training wall
Google fires back today and says we're
not hitting a pre-training wall that's
only if you don't have
imagination and we we will have to see
how that all plays out I want to call
out something that I think people
overlooked Ilia has had opinions in the
past about what's just ahead they're not
is correct especially in terms of
sequencing it's very difficult to get
the future right sequentially it's
easier to get it right conceptually if
you deeply know your field he deeply
knows the field and so if you look at
past statements about how large language
models will progress he's often
conceptually
correct
but he has been so hesitant about what's
ahead and that came out at NPS it's like
this moment past with Chad GPT and
reaching gp4 level now 01 and the sense
that nobody really talked about that
that I got from Ila's
talk is that he's not sure what's ahead
for the first time in a long time and I
think that's worth talking about that's
a
story and he he hazarded some guesses
about how we get to Super intelligence
but for the founder of safe super
intelligence to really publicly not be
sure about the next step forward is a
remarkable
thing from the money graph on his
presentation I think his bet is on
recursive AI self-improvement but that's
very
unproven and the money graph by the way
is this like line chart where he
basically calls out that
intelligence and brain size in mammals
are somewhat correlated except for
hominids us and we have found a way to
hack our wetwear hack our brains so that
we are able to have intelligence that
scales that isn't linearly linked to
brain
size great love that it's a very astute
observation but it's an analogy it's not
exactly clear how that would map to
artificial intelligence and Ilia made it
pretty clear that he's not sure yet
either now it is possible he is
concealing Alpha right he has some
special Insight on this that is the
founder of safe super intelligence he's
going to keep to himself until he can
prove it that's a possibility
but my sense is he's actually trying to
figure out the way forward with the rest
of
us and that's going to make for a very
interesting 2025 because there's a lot
of candidates there's people who are
saying extra test time extra inference
time when you type in your chat is going
to help us really scale intelligence
there's people saying synthetic data is
the way forward and we don't need an
internet's worth of data for pre-t
trading because the llms can generate
their data now there's people who are
saying that it's about logic which is
something that ilot tipped his cap to in
the presentation saying that if we can
teach machines to be logical we have a
path
forward this feels like a hinge moment
it feels like the beginning of 2025 is
one of those inflection points where
there's multiple paths into the future
and some of the leading model makers and
some of the leading thinkers in the
space are still trying to figure out
what is a path toward the next step
change in intelligence and what does it
look
like we shall see I'm excited to find
out but those are things that I think
are more strategic versus the open AI
search stuff which made headlines I hope
you enjoyed this cheers