andxor

I'm not defending Dario. That's not the point I'm making.

Yeah, I think the answer is somewhere in between. Anthropic has been a few months ahead of everybody in terms of internal capabilities.

So it's not 6 months but it's also not a few weeks.

Normally I would say release date to release date, but Mythos/Fable release was significantly delayed because of security concerns.

Mythos/Fable was trained 6 months ago.

They are actually paying them $1B per month for the compute. But they're still profitable and increasingly so.

What makes you think the API costs are "the true price" and anywhere near their inference costs? Also what percentage of users do you think actually maxes out their subscription?

I suspect they're subsidizing a lot less than people think.

Exciting! In my wildest dreams I'd love to write an OS with this.

You shouldn't take financial advice from me :)

Exactly. As usual people are bullish at the top and bearish at the bottom.

That's not what is implied. Fable may still be widely available.

I used Fable for like what 2-3 days at most and didn't really feel it was so much better

It was a lot better. I can't believe people say this.

This page has existed since way before the current Fable controversy.

https://x.com/trq212/status/2068793885535694858

I'm sorry.

Another opinion piece?

This is an opinion piece.

I have had the same experience. I can't believe that people couldn't tell the difference.

I'm calling out the "tech opinionist community" broadly speaking. I don't know if it's the same exact people but it's often the same forums or news outlets.

The same people who told you Google was going to zero three years ago because Search would become irrelevant are now telling you users will move away from Google because it’s removing Search.

When will people learn that the quality of Google’s leadership is better than that of the average Joe on Hacker News or a random tech journalist writing clickbait articles?

Codex is also opensource.

Then the benchmarks are wrong.

You're so naive. It's all a big game of domino.

Hyperliquid.

Great gonna add these to my CLAUDE.md /s

Luckily Germany is not an emerging market :)

I think it's pretty obvious from their financial results that this company is a commercial failure, and the subset of people who consume their product on a regular basis is vanishingly small.

Do some people occasionally eat fake meat? Sure. Enough to build a sustainable business? Less clear.

Arch Linux wiki is the gold standard and better than FreeBSD.

Your usual reminder not to come to HN for macroeconomic analysis.

You're not doing anything wrong. The Chinese models are not as good as advertised. Surprise surprise!

The models are hardly equivalent.

Oversight by whom?