Jeff Bezos creates A.I. startup where he will be co-chief executive | Hacker News Reader

Jeff Bezos creates A.I. startup where he will be co-chief executive

nytimes.com

170 pointsdominikposmyk145 comments

Wouldn't be surprised if the funding is based on a circular investment of some AI chipset vendor to consume the compute power of the chipsets AWS is buying.

"We buy this amount of chips at price xx, you guarantee us utilization in part by investing in an AI company we will create"

Why does it feel like the entire US economy at the moment is supported by 5-6 companies constantly moving around a couple of trillion dollars?

Because we’re in a bubble.

People have been saying we are in a bubble since before covid...

Not sure if that makes the situation better or way worse.

You know can bubbles grow, right?

Which means a bigger burst.

Because it is

And for his sin of stealing fire from the gods he’ll be chained to a rock and an eagle will eat his liver every day for the rest of eternity.

For the humor, mocked up an Amazon page reflecting this scenario:

- - -

Human Liver, 14.4 cm

Visit Jeff Bezos' Store

3 Stars (1)

$14,000,000

Coupon: [ ] Save an extra 5% on your first Subscribe and Save order.

In Stock

Quantity: 1

( Subscribe )

Save 5% now and up to 15% on future deliveries

SNAP EBT available

Delivery every: 1 day (Most common)

The UN doesn’t say that a human life is actually worth $10 million but that’s the number we’re going to use, so the question is as QALY thing how much is a liver transplant worth and it turns out the liver transplant gets you about 25 QALYs, and so running the numbers a liver is only about $3.75 million.

> The new company has until now kept a low profile, and when it was started is not even clear.

Any more details? Where is it located? Who is working there,

For $6.2 billion raised I’m surprised their aren’t more details

Is that a thing in the US? You start a company and there's no need to register it with the government? Or it gets registered but there's no public records of it?

You can basically form a corporate entity with a nominal Delaware office, but it doesn't need to give any details about where the actual work takes place, yeah.

You can even LegalZoom one for $200.

All companies are registered. They have to be registered to be legal entities, have bank accounts, and comply with tax laws.

Private companies don’t need to publicly divulge a lot, though. It’s between the company and their investors. It’s only once a company wants to trade publicly that they have to provide a lot of public details and financials.

You don’t have to register a general partnership as long as it has one of the partners’ last name in the partnership name, although I guess you have to get an EIN to file partnership taxes.

A sole proprietorship doesn’t have to register anything ever at all.

Great point. I suppose I should have said all companies like this (the corporation Bezos is involved with) are registered entities.

There are ways to do business activities yourself without registering an official business, though it’s generally discouraged because forming an LLC is so cheap and easy and provides some protections and benefits.

Although in most cases it’s sensible to register a single member LLC instead of operating as a sole proprietor. That way the LLC can be separate from the owner’s personal assets.

There are thousands of companies registered every day across the US. This one is probably a subdivision of a subdivision of some holding company owned by Bezos. Pretty much impossible to track using just public data.

In my state, zero information is given to the state about who the owners are.

Let me guess: Texas?

Nope. TX actually has quite a bit more transparency.

One of the more absurd things I can do is have two LLCs own each other, and then have outside management.

Wyoming?

A company can be a sole proprietership or partnership without registering with anyone, but in this case in most states they must either conduct business in the owners/partners names or register a DBA with the state(s) they are doing business in. (the rules are state specific, and I think there are some states that don't do kr require DBA registration). In most cases, a company with billions invested in it will be formed as a formal entity such as an LLC or Corporation in a state. Again, the specifics vary from state to state. If you knew the legal name of this entity, and what syate it was registered in, you could probably look up when it was registered in that state.

However, details like owners and organizers aren't always Available.

It gets further complicated with Series LLCs.

Congress passed a law that would have required "beneficial ownership" registration with law enforcement (FinCen), however, this registration would not have been public.

Further, it was found unconstitutional and enforcement of the registration requirement indefinitely suspended.

In general, if you are doing business in a state under a name or entiry other than your own legal name, you will be required to file something with the state, and that filing will include a registered agent where legal process can be served on the business, and this information will be public.

But if they aren't doing business publicly yet, no one will know the name of the business, so they can't look it up! It sounds like the name mentioned in the article may just be a code name.

Interesting. If one were to legally change their name to their desired DBA name, I wonder how that would go over.

Legal name changes are handled in state courts, so there may be different rules depending on your state, but generally I think you could get away with it if you really wanted to... but doing business as a sole proprietor means you are personally liable for everything the business does. Way simpler and better to just form an LLC in most cases.

$6.2 billion which are prob to be invested in amazon to be invested in the startup to be invested in amazon, so I would not assume such an amount actually implies the size where this would be surprising.

More circular financing to keep the AI house of cards from falling?

Yeah since there isn't any actual real spend going into AI...

I don't think many of us are questioning the real spend going into AI - it's the return on that spend that we're wondering about.

There is pretty concrete published data that shows trends in less hiring for roles that are easily replaced by AI such as content writers and front-end developers

Personally I let go a developer I was using because I only had them around for front end work and now I'm much more productive just doing that with Claude Code directly

So the returns for the average business are largely due to less employee and contractors spending

> There is pretty concrete published data that shows trends in less hiring

Is there any concrete evidence that lower hiring is due to AI, and not due to some other factor - such as a stalling economy? I suspect AI growth is the only thing currently staving off a full-blown recession.

I alluded to it in my comment but let me make it even more clear

There is significantly less hiring for roles that AI can replace easily compared to roles that it can't replace easily

If there was decreased hiring simply because of the economy then why would it only be impacting certain job roles? Hmm... Mystery...

Have you given thought to the fact that there is an overlap between "jobs easily replaced by AI" and "jobs that are easy to consolidate into existing roles", or even "non-essential jobs"?

> jobs that are easy to consolidate into existing roles

I think you are on to something here. AI is what gives people the ability to consolidate many roles into one.

Like in my example, when I fired my front-end developer, I now can easily do that task.

Or, a marketing generalist can now create blog posts using ChatGPT instead of needing to also have a content writer

To your point about "non-essential": "Front-end development" is not necessarily something I would consider non-essential, but maybe "front-end developer" is now. Something to think about

If you're vibe coding a front end, you're going to run into issues with technical debt eventually, so that's a very short sighted way to handle things.

Or if you do have the technical skill for front-end work, you could have done it without the AI.

I do doubt this, however, because AI assisted programming does not increase productivity for skilled workers as much according to studies.

I've been programming for well over a decade, but as a founder I wear many different hats so I had a front end developer to help take some of the load off.

But AI coding got to the point where, for me personally, it took less of my own time and effort to work directly with the AI (Claude Code) to do the front-end tasks required compared to working with the developer, reviewing their code, etc.

I'm not "vibe coding" or adding technical debt.

> you could have done it without the AI.

Technically anyone could do anything, but there are finite things in this world like the number of hours in a day lol

> according to studies.

Great, I don't care about the "studies". I'm talking about how things work for me personally in the real world and giving a concrete example of how AI spend has replaced an employee but you are welcome to ignore that

Also, it should be noted that in studies, the conclusions are typically based on an average or median, so some people will see a benefit, and some people won't... and that will be based on a number of factors

There is money going in, but not as much as what these valuations and investments imply. The numbers are inflated and that's common knowledge by this point.

> that's common knowledge by this point.

No it's not. Go short the AI companies though if that's what you think

The circular deals going on are documented. There is no point in discussing if this happens or not. The numbers just cannot be added up.

Now if things will fail, how and when is a different discussion.

Two things can be true

Yes there are circular deals going

Yes there is actual underlying spending from the average consumer and business that makes these valuations largely realistic

Q: could this be an "experiment" in AI financing? i.e. he's bankrolled 6.2B, then strategic quasi-purchases and cross investments will multiply that money without ever needing to spend it on anything - except AWS hosting of course

based on linkedin it looks like they took a failed company thats existed for a year "General agents" which had what just looked like a single browser agent and now have rebranded that.

From the TechCrunch article;

> its work will resemble that of Periodic Labs, which is building technology to speed up scientific research by simulating the physical world to train AI models.

Will be interesting to see how far simulation gets you vs actual embodiment via robots, etc.

Everyone knows that the best companies have two CEOs.

Seems to be working for Waymo so far, though of course Waymo isn't an independent company.

Despite the critiques that is something worthwhile I can understand, maybe there is a time in your life you want to be involved in something big, but not 100% like you were at your prime you have the opportunity to do it, so why not. You remain engaged. I prefer seeing that than doing nothing of something useless.

At this point I prefer these big tech CEOs just be involved in whatever is least damaging to society.

I think that’s a very mature, realistic take. Personally, if I was a 60 year old billionaire, I’d probably be having a series of long vacations or chilling in my yacht. It’s cool that bezos is still trying.

Maybe you wouldn't. Maybe you'd find it difficult to walk away from the lifestyle that made you a billionaire in the first place.

Maybe! I certainly hope so. But if my life thus far is any indication, I do enjoy working hard but also have to fight _really_ hard to not slip into laziness, even though it makes me less happy. Something about the delayed gratification of doing hard things -- there is a ton of gratification, but my monkey brain hates the delay.

A lot of negative posts here, but AI to advance science seems like basically the best possible use case. The more ultrawealthy people who want to throw billions at it, the better.

It depends on how much they're actually doing in the service of science, and how much is "flashy AI stuff".

I don't have a lot of hope that it's the former, to be honest. These people have burned up all their goodwill.

Wasn't OpenAI also for the embetterment of society? Looking at how things went for all high profile AI companies, why should we trust this one? Even if they got a saint to run as CEO they would find a way to screw humanity over.

Notably, Amazon has already invested $8 billion in Anthropic/Claude, so I'm hoping this is actually something wildly different and with a different approach.

Well, at least it won’t be a fake nonprofit… I hope.

Well, Elon gets enormous amounts of tax dollars enabling him to move resources around between his investments. The billionaires aren't differentiable.

> The billionaires aren't differentiable

Probably because they're not continuous?

I feel like he's too distracted by his celebrity lifestyle these days to really focus on running a company.

He just paid for Kris Jenner's 70th birthday party at his house where they had the cops called on them: https://www.realtor.com/news/celebrity-real-estate/kris-jenn...

and he disinvited Elon Musk (lol): https://www.yahoo.com/entertainment/celebrity/articles/elon-...

Here he is out showboating at the Oscars, Vanity Fair parties, the white house, in Hollywood, in Monaco, Paris Fashion Week, Sun Valley, Milan, NYC, various galas all in the last year: https://www.businessinsider.com/jeff-bezos-lauren-sanchez-fa...

Pretty gossipy

I don't want to be mean, but to be honest, your post makes me wonder if you are too distracted by the celebrity lifestyle (of others).

I don't have this treasure trove of information about Bezos' life and I don't think it makes me any less informed about the world?

Haha.. busted. I do enjoy fashion blogs (specifically Tom and Lorenzo https://tomandlorenzo.com/) and always notice Bezos out galavanting around at the events they post!

Being a celebrity doesn't sound like it would take that much time, especially for someone like him who probably just is physically present for things and leaves all planning to others. In fact, being coCEO is probably literally just more of the same to him.

You seem very up to date with all of the celebrity gossip.

I have a weird sense that this is a way to get his kids that are technical into a family business without having them work at a company that isn't considered prestigious

Is "A.I." NYT house style? Looks rather jarring.

> Is "A.I." NYT house style? Looks rather jarring.

I think their style is to use periods for acronyms, which I believe is traditional. A quick scan of their recent headlines turns up "U.S", "A.I.", "A.T.M.", "ICE", "L.P.G.A.", "REI", "U.K." I don't know what the reasoning behind the use of "ICE" and "REI" is, could be a mistake or a judgement that those words are tend to not be understood as acronyms, or something else.

IIRC they used to always style "NASA" as "N.A.S.A." even though the agency itself never uses periods and is of course always pronounced as a word rather than initials. (This particular example stuck in my mind just because I work there). Hopefully "ICE" and "REI" reflect a change in that style to omit periods when referring to organizations that omit the periods in their own style guides.

I don’t think this has been true for the vast majority of NASA’s existence. [This 1966 NYT article about Gemini](https://www.nytimes.com/1966/03/20/archives/gemini-8-mishap-...), [this 1986 article about Challenger](https://www.nytimes.com/1986/09/22/us/nasa-had-solution-to-k...), and [this 2010 article about Constellation](https://www.nytimes.com/2010/05/27/science/space/27nasa.html) all omit periods. Even if NYT did stylize it as “N.A.S.A.” in its earliest articles, this would not diverge from common agency practice early on. For instance, in [this official reel about the dedication of Marshall Space Flight Center](https://www.youtube.com/watch?v=4fasslw4MvE), it is pronounced spelled-out as four letters.

Looks to me like initialisms get periods unless they are acronyms or trademarks. So "ICE" is fine because it is read as "ice" and not I.C.E. (eye-see-ee). REI is not an initialism though, but presumably they have kept it as-is because it's their trademark/"doing business as" style.

That is correct and also how everyone else should be using the English language.

The “REI” initialism, is the official d.b.a. for the Recreational Equipment, Inc. company.

Erratum: I meant to say that REI is an initialism but not an acronym

Artificialïntelligence

This is the NYT, not the New Yorker

Recap for those before the paywall:

New AI co, Bezos as co-CEO (alongside Vik Bajaj, ex-Google X)

$6.2b in funding

Nearly 100 employees

AI + real world scientific experiments, for the engineering and manufacturing of computers, automobiles and spacecraft

In what world is that still a startup!?

I thought startup was supposed to mean "we're still starting up our business", but it sounds like the meaning these days is closer to "we are setting piles of cash on fire".

https://archive.is/ZhmTd

Engineers were modeled after him?

I read "co-defendant" for a second there. Oh well.

Everything from A to I

Jeff Bezos as a CO-CEO sounds like a corporate structure recipe for disaster.

Why?

How do they break ties?

Dip said ties in liquid nitrogen and hit them with a hammer

It’s pretty common sentiment with senior level Amazonians that Uncle Jeff was a great leader. Many lament his departure and loathe his replacement.

Lifestyle distractions aside, Bezos playing any kind of CEO role is probably a good thing.

I wonder what the dynamic between this and the Amazon Alexa team will look like, if there will be any in the first place...

I think it's actually going to be healthy for the ecosystem on the whole. The more competition, the better.

If you believe that rushing capabilities ASAP is the right call, that is.

He saw all of the grifting at OpenAI and said to himself: "I can do that!"

I don't get the downvotes. This is a case of the simplest explanation being the most likely one.

Just before the AI bubble goes pop too. I think by my estimates there's three more months left...

Not many people can do the "Just-do-what-Musk-does" playbook, but this will no doubt be a good little bunse for him.

If there were a founder I believed who could, it would be Jeff B

The fact that Amazon hit it big twice under him (retail and then web services) speaks volumes. Hard to pretend he just got lucky.

YAAS (Yet Another AI Startup). Yawn.

In the fires of Mount Doom, a 6th frontier lab was secretly forged. For none could resist being a CEO of an AI startup before the music stops.

When you’re rich, you can even buy vanity CEO roles.

Openai, deepmind, anthropic, xai. who is the fifth frontier lab?

FAIR (I mean MSI) or grab bag of other runner ups

Microsoft Advanced AI Research

> Co-Chief Executive

In other words he wants to seagull manage the place.

Wants all the power without actually doing the CEO's job. Quite ridiculous, similarly how there were two heads of "DOGE" or Twitter having Linda Yaccarino as the CEO.

Linda had an important role as the Chief Bag Holder in case things went bad.

Prometheus? Eagles > seagulls...

"I bring the money, so I have the last word on everything. Now talk."

Co-CEO. So he gets to boast about being CEO of an AI company at dinner parties while leaving someone else to do the actual work.

> CEO

> actual work

Doesn't compute.

Value I think is debatable. But most every CEO I have met is the workaholic type.

Doing work isn't necesearily value, and value depends on perspective.

Like I said, value is debatable.

I’ve met both. Even when I disagree with them I appreciate the ones that actually put the work in. Most recently I’ve worked with a string of them that barely understand how their companies make money and certainly couldn’t do any of the actual jobs there. Performance is independent from them being on the payroll.

The Amazon Prime Video approach

> Called Project Prometheus

What is it about tech people and being unable to come up with original names?

Every company I have worked for has had two dozen internal tools and projects called "Prometheus".

setting aside the AWS DB offerring I've done 2 "Project: Aurora" in the past 3 years, and a bonus project "Audite" (make sure you say it correctly when execs are around!)

Look, I'm all for "big bets" but when the majority of time and effort goes into the naming, kick-off and t-shirt design, this ain't it.

> Project “Audite”

Reminds me of Matt Levine joking about walking into a hedge fund with a SEC jacket memorabilia.

https://archive.ph/K1sNB

So we now have one AI startup run by Jeff Bezos and another AI startup run by Beff Jezos. What a weird timeline.

Expect a third soon. Bezos really wants to have another hit.

I think it’s great that there will be more competition in the space, not to mention more hiring, etc.

I wasn't referring to Amazon. There's actually another person called "Beff Jezos" who runs an AI startup called Extropic.

Amazon itself has become so corrupted and disfunctional that all of their internal AI efforts amount to just burning billions for bottom-of-the-barrel results[0]. I'm not surprised that 5 person startups are building better AI products than all of Amazon, or that Bezos decided to start an AI company outside of Amazon.

[0]https://labs.amazon.science/

It's really not surprising. AWS is definitely absolutely now in the category of "too big to fail". And the fact that there's an entire AWS region for US government classified stuff just underscores that.

Becoming TBTF has a very corrosive effect on organizations.

I wasn't referring to Amazon as the other AI startup. I was referring to another guy who is known by the name "Beff Jezos", who runs Extropic.

Oh! That is indeed strange. I bet Beff gets some interesting emails…

Real name Guillaume Verdon

Honestly, why? Why not just focus on Blue Origin?

I don't get this new Musk-like tendency to run multiple ventures instead of focusing on one very tough mission.

It takes extraordinary skill to successfully juggle multiple ventures. So it is natural for some to want to take on the challenge. I think it is pretty impressive.

Maybe he has this skill, but I also know a lot of folks who like to have the control and glory, but don't want to put in the time basically.

And he's not juggling Blue Origin that well given the delays. Still impressive to go to space, so the man deserves credit, but it likely would have been several years faster and billions less if he had been more involved.

By pure nature of how companies work, a space company with this mandate and so much funding, unless its being used for money laundering, will have a modicum of progress. BO has barely had that. No space company with so much money and so much runway has achieved so little.

> It takes extraordinary skill to successfully juggle multiple ventures.

That's a myth. I've done that, and I know a lot of people who do that. Do you think Musk is writing sparse attention code for Grok? Does he even know how Grok's architecture works under the hood? Or that he designed the data centers? I mean, you delegate stuff. The only hard thing is getting the right people, but if you're a hyped up billionaire, it's easy mode because you can pay a lot, and people want to work for you. You just create an environment where they can achieve things.

There are times when the majority of your work is simply attending public meetings, podcasts, and doing interviews. People really overestimate what's involved in the work of a billionaire CEO. The people actually making things happen in space industry or AI work harder, longer, and solve more complex problems than any CEO and in some cases they need to work hard against the CEOs to actually make things happen.

> if you're a hyped up billionaire, it's easy mode because you can pay a lot, and people want to work for you

blue origin vs spacex says otherwise..

Blue origin just landed their new glen rocket. Not an easy feat.

Although I wish billionaires would fix homelessness, I think its good there’s more competition in the space launch industry.

Homeless is mostly a self inflicted issue that most don’t want to solve. So why should others try solve it?

In the old days there was a saying "The only way Bill Gates could spend all his fortune is a manned mission to Mars."

Well, Bezos and Musk are trying that, and it turns out there is still a lot left to spend, so instead of helping the needy they do what is trendy for billionaires, which is build another AI company.

Once they have the AI company, thennnn they will help the poor right? They will use AI to help the poor!!!

It’s free money. You spend 3billion and have more billions drop in from investors. You get to create great tech and then make 100s more billions drop from IPO. When the initial investment is a rounding error in your portfolio, it’s a crime to not do this

Or they could just be happy with the many billions they have.

Realistically if they were capable of being happy doing something other than acquiring more currency they would never become billionaires. Any reasonable person would simply live life enjoying their 10s or 100s of millions.

A lot of people unfortunately believe in the musical chairs theory of markets.

It’s a simple way to sound cynical and savvy. Good thing it’s also extremely incorrect.

Elon's comment: https://x.com/elonmusk/status/1990479697793491195

[0]https://labs.amazon.science/

Reply on news.ycombinator.com