epsteingpt

So many books, so few people to read them!

Sneakerbots killed it for everyone.

Priced out the demand IMO too early.

But also, people only have 2 feet.

AI isn't showing up in S&P 500 company earnings as a line item. proof: https://aiearnings.dev

given the stakes, we'd better hope it starts showing up soon.

This is a good headline and point.

Realistically, they're worse than a central bank, because they can't exactly expand supply monotonically like a normal central bank. Nor do they realistically control rates.

The bioweapons threat is real, but shouldn't be constrained at the 'intelligence' level. Rather it should be constrained at the supply chain level.

The cybersecurity threat will likely be a cat and mouse react game for a while. Just like robberies / the mob was in the early 20th century.

Social forces bring things into balance over time, much more so than the proactive actions of individuals.

Unfortunately, the genie at this point is unlikely to go back into the bottle. There's enough 'intelligence' out there that a super intelligent model could emerge at some point in spite of pacing.

This isn't a doomer scenario - we tend to navigate social changes better than we ever could have hoped.

Shocking this isn't bigger news. It will be strange what they get you for. Maybe car ownership isn't doomed after all.

At least a human usually has some judgment and discretion situationally.

I guess the alternative "Juvenile Driveby Shooting takes place from Waymo" isn't a better headline though.

Bro if these models can break into hugging face, they FOR SURE can break into OAI's internal databases and change a flag.

Humor is too often lost on HN

Audit is the way. Even better is having one model family like Claude check another model family like Codex.

What will be even more shocking is the future AI agents in clinics saying:

"Oops, you're totally right. That particular gene interaction is fatal. Would you like me to repair the gene therapy to account for this?"

Such is the true cost of progress.

Only here to say, regardless of the drama, shouldn't we all be excited if the Navier-Stokes gap is closed?

Time will almost certainly reveal a lot more about the drama and the related ethics, but let's get excited about the actual breakthrough as well!

so many web conventions that 'don't make sense' are about to become exploits.

hilarious.

For all the smarm and jokes, IBM still makes ~$70B in annual revenue and enables billions more in systems integrators. Not many companies can say that.

This allows them to take the secondary revenue in-house and potentially fund a new mainframe business.

It's a [smart-ish because it's their only] strategy for the near term, because long term businesses will be moving off their mainframes and onto more powerful and intelligent systems that can 'think'.

Business Machines 1.0 were CRUD; if they pull this off they can at least get enough capital to maybe upgrade business machines to 2.0

Good on them for at least trying.

It absolutely is.

1 hour minimum wage vs Big Mac price in the US :

1980: $3.10 vs $0.50

2022: $7.25 vs $8.00

That's the whole thing.

DHH mentioned the fourth turning as an inspiration about cyclical history. With LLMs historical analysis is simple, this webpage is the result. Some slop, but overall a nice primer. Enjoy.

Sad, but this deplatforming nonsense was always going to cut both ways.

The slippery slope continues.

Robotics data remains scarce and expensive, leaving many important tasks outside the pre-training distribution. Out-of-distribution ICL is therefore essential for general-purpose robot policies. S1 ICL can drive ultimate success in novel tasks 2/3rds of the time.

Nice idea.

Related: https://news.ycombinator.com/item?id=49432761

AI needs to show profits in the 'normal' sector to grow into its earnings and valuation.

nothing has happened yet, so it could get ugly pretty soon if "normal" companies don't start seeing P&L

Very interesting. related: https://news.ycombinator.com/item?id=49432761

What's most interesting is that this is driven by the AI boom as the tech companies themselves secure and drive circular earnings that have not yet paid off yet.

NVidia sells chips to broadcom which books capacity commitments from Meta who then invests again in NVidia to develop its own chips.

This isn't bad - seller financing works in cars.

The issue is that they need to sell more stuff (ads, products, AI) to 'normal consumers' eventually.

It's going to be an interesting 2027 and 2028!

Despite flashy AI-boom revenue growth at labs and chip companies, AI companies ultimately need to drive profits in "normal" companies to grow their revenues and into their elevated valuations. I orchestrated AI to deconstruct the 10Q / 10K filings for the past 30 days, and found that while most mention AI, almost none outside of AI suppliers show concrete profit driven by cost reduction, improved service, or new business lines.

China Bros - here is visual PROOF of the Tang dynasty.

https://en.wikipedia.org/wiki/Tang_%28drink_mix%29

Many ancient flavors and notoriety was achieved in that dynasty. Mango Tang was my favorite.

One of the most powerful slides on the topic is Ari Grant's

He showed two pictures at a frustrating situation. The first WHY is this happening? Shaking a fist at the sky in anger. The second why IS this happening? Curiosity.

The second leads to revelations and growth about the company. It also doesn't endanger your job or make people want to work with you.

Work is a social game. Best to learn its rules rather than rage against the machine.

They're really trying to tank Anthropic's IPO huh?

the sentiment is negative and justified. anthropic nanny states what you can do.

in your instance, anthropic may decide, arbitrarily, to stop 'autonomous rebuilds / refactors and ports' because they could pose some alignment/rights/etc risk to whatever slop their philosophers dream up while they're out eating $200 avocado toasts. then you can't do the thing anymore.

fable is good, absolutely. agree it roasts Sol which is, comparatively, a little receipt-hunting jack**

but now imagine being an enterprise, and having another organization not only taking your workflows and baking it into your models, but then deciding they can arbitrarily cut you off.

when you can instead own your data, use an agnostic provider, and get better results (through model combinations), it will take 1-2 quarters to figure it out.

the main reason anthropic is killing it is because they really do understand the enterprise development experience and lifecycle and have built products and have a sales-team that can deliver.

business-model and vibes-wise they have lost all goodwill in the past 6 months, and that momentum will be quite hard to regain.

A depressing finding. The unfortunate reality seems to be the 'market' is pretty small for these hotlines.

We need to solve the underlying issues of loneliness

The Romans didn't have the Federal Reserve.

Correct - no one can stop the fiscal train!

No one serious is worried about American solvency. The paper says 50% over the next 10 years, but even most economists misunderstand how the monetary system works.

There are so many other issues to worry about at the moment more immediate than solvency.

Getting lost isn't fun. I resent the people who say so.

In many places in the world, a missed exit on a freeway can take you 20-30 minutes or more from your destination, with no chance of turning around.

There is an art to wandering, for sure, but this luddite paean to wandering neglects the fact that for most of human history, one of the largest drivers of technology was navigation technology to AVOID getting lost on purpose.

This is projected metaphysical angst masquerading as nostalgia.