chongli

Close, it's the time-discounted expectation of future returns. This seems related to future profits but it need not be. Historically, stocks tend to perform poorly after IPOs. There's no guarantee that (say) Anthropic's stock price would ever recover after a post-IPO drop.

The recent attempts by Anthropic et al. to circumvent the usual rules for inclusion in indices have raised red flags all over the place, with many calling it a naked attempt to raid everyone's pension funds for hundreds of billions in ill-gotten capital.

The same could have been said (and has been said) about other tech company employees for all sorts of other reasons in alignment with those companies' goals. Don't you remember how much people used to laugh at Tesla employees for worshipping Elon Musk as some kind of god of engineering and entrepreneurial genius? Or Apple employees in the Steve Jobs reality distortion field?

I would have thought at this point that it'd be well known that the employees of all cult-like tech companies can't be trusted to make a sober evaluation of their companies' justified valuation. We can talk about conflicts of interest and we'd barely be getting started! How about biased selection by hiring managers for the most fervent believers in the company's mission from the get-go?

The question is not whether AI is useful. It is useful, full stop.

The question is whether AI is useful enough to justify valuations that dwarf the GDP of all but the top 20 countries in the world. As it stands right now, it's not even close! The leaked OpenAI financials put these AI companies in the same profitability territory as utilities, with zero justification for these crazy valuations.

The whole point of the article is to answer this question, and here's the answer:

Because all the AI doom fear-mongering is driving sky-high valuations. The more the public panics, the more investors open their wallets.

If you have to interview more than a handful of candidates to hire for a non-specialist role then something is seriously broken in either your hiring system or hiring for those positions in general.

That's not true in all RTSes. Take StarCraft, for example, and there are plenty of games on record that were decided not just by 1 unit, but by 1 attack from 1 unit. There are Zerg players, for example, who have developed a reputation for creating havoc after getting a single zergling (the smallest and cheapest attacking unit) into their opponent's base. A single shot from a Protoss reaver can mean the difference between taking minimal damage and losing half of your workers (and subsequently the game).

Channeling is usually caused by too fine of a grind. If your machine (I'm assuming it's a pump machine) is pegging the pressure gauge at max (and dumping excess pressure internally) and your coffee tastes unevenly extracted, you may want to try grinding coarser. Not only will this reduce channeling, it'll result in less fines in the cup, also reducing bitterness.

The best thing I ever did for my espresso was to give up on the rigid rules I was first taught as a beginner. I don't time my shots, I don't use fixed brew ratios, I do everything by feel (watching the pressure build and the coffee flow) and taste. I do use a scale (for weighing beans per dose and weighing shots for repeatability). I dial in by adjusting the coffee output rather than fiddling with the grind. I only set the grind once to get a reasonable pressure (6-9 bars, no maxing out or dropping off), then fine-tune the gram output.

The biggest insight I gained from this freestyle approach is that the standard 2:1 ratio is altogether wrong for most of the light-roasted coffees you get from specialty coffee roasters. They simply will not extract properly with that small amount of water. Grinding coarser and pulling a longer shot (sometimes called a "turbo shot") gives you a much better result.

That rests on 2 assumptions:

1) That inference on OpenAI's frontier models is actually cost competitive with open models. Their high SG&A suggests otherwise.

2) That slashing R&D won't lead to a marketshare collapse when everyone (remaining) moves to Anthropic to get on their frontier models. All evidence suggests otherwise again, with Anthropic already exerting enormous competitive pressure on OpenAI's marketshare.

I think OpenAI is in a terribly tenuous position: they're getting squeezed from Anthropic (on the high end) and open models on the low end. A lot of companies in a lot of industries suffered this fate. Getting stuck in the middle is not a good thing!

Nature abhors a vacuum and that principle extends to power vacuums.

You're not going to take a 60% paycut if it means 60% less food, 60% less shelter, etc.

The claim that the North Korean dystopian dictatorship could be generalized to all cultures, across all cultures, merely on the economic and military capabilities of AI, is an extraordinary one. It relies on a great many assumptions about the political as well as independent, personal, and organized responses to the societal changes that would need to take place in order to bring it about.

A sloop’s guns were generally so light the balls would bounce off the hull of a ship of the line. The sheer weight of firepower in return would shred a sloop to pieces. Add on to that the fact that heavier guns can achieve longer range when trained at the right angle, and a sloop can’t really get anywhere near a ship of the line and live to tell the tale.

"Over a six-year period between 2019 and 2025, almost half of a large set of denied health insurance claims in New York state were reversed when the cases reached independent review organizations, comprised of clinicians unaffiliated with insurers"

But that doesn't get into the reasons why the claims were denied in the first place. It doesn't tell us anything about bona fide spurious denials vs. improperly filed claims (mistakes in the paperwork), clerical errors, or clients placed under investigation for claiming too early (after applying) or too often (making a lot of spurious claims), or care providers who do the same.

Insurance companies are concerned with adverse selection and moral hazard. A client who files a lot of claims shortly after getting insurance raises the suspicion that they were not honest about their health prior to applying. Similarly, a client who claims every drug a pharmacy carries raises other suspicions.

Of course, most clients aren't like that, but a not-insignificant minority are, and a small number of clients can file a very large number of claims.

Spurious denials? Or improperly filed claims?

I think that's the point. "How do I say they're lying without outright saying they're lying?"

It's a common rhetorical trick.

2^2.2 = 4.595, 255^2.2 = 196,964.699

What could be better than that?

It should be considered common courtesy that when you ask a question you have at least attempted a bit of research to find the answer on your own. Then you can explain why your attempt to Google for the answer failed.

Of course that may be breaking down, as search engine results quality has declined dramatically in recent years.

It doesn’t. Computer interfaces had no superfluous subservient text for their entire history prior to LLMs. Some of these interfaces have been highly efficient as tools, arguably more efficient than more recent software in many cases.

When people complain about LLMs being subservient, they’re not complaining about the tool fulfilling their request. They’re complaining about being forced to read a lot of superfluous, overly polite, or even self-deprecating language. There’s nothing in the entire history of tools (going back to Neolithic times) that would indicate that we need that. All of that stuff is an artifact of social interaction between humans in the presence of cultural norms.

When you’re alone in your shop with your tools, you don’t need your bandsaw to apologize to you for nicking your finger.