sroussey

Can’t buy and ship those.

I love how ASML machines are shaped to the contours of a jet fuselage for this reason!

I’d say Japan is a more important source than sources in Europe.

HuggingFace has been extremely responsive when an issue is raised. I found issues in their infrastructure this year as well as cache confusion in Chrome, and HF fixed their CDN in a day and I didn’t hear back from Google for a week or two and could no longer show how to exploit Chrome as I figured HF infrastructure issues would drag on possibly forever.

Negative margins to have you hooked on someone else’s software?

It’s a test with a good excuse. If caught, it’s only their own data, right? If not caught, then time to abuse the system.

Carry propagation is indeed quite slow. With so much math for AI, is it finally time to go back to analog? Superposition is instant.

So the code for the web, the desktop app, and the cli where largely similar.

Try prompts that convert the gif to svg? Maybe the models should do that internally… start with their image model first and then make an svg.

Once hired a DBA that reworked backup scripts. He got real annoyed at the idea of testing the backups with real restores and clearly never did on the real databases (only on his smaller samples).

Turns out the backups took 30 hours. The daily backups. That then overwrote each other on the assumption that the backups would not take that long.

Of course we found out the hard way.

Or they are sending different code to different people from the same url…

Or a spaceTwtterAi stock holder…

How much of the speed-up is attributed to not hardening the code?

And i do not mean this in a flippant way, as how to harden with speed in mind might alter how to design the format and the codec.

You mean bylaw not law. No law says anything about super voting shares not surviving death.

Um, 11 months old is pretty recent..

No, that misses the whole point of content negotiation (for language, et al). In the age of search engines though, different urls are often perferable.

You definitely are required to disclose what frequencies are used and at what power.

It’s terrible and depressing work to take vibe coded garbage and make it a real product. There will be demand, but good engineers won’t want to touch it. And people paying will think they did the hard work so why pay a good rate?

We used to call this snow shoeing.

It’s almost like some people do want to

Elon is known for purposefully not naming people correctly…

Agree. But just because it feels the same doesn’t mean it compiles the same.

There are a lot of search APIs for AI these days.

/s

You can write a skill (many have) to lead it in how to use different models and efforts for subtasks. For searching the code base, for example, I have it use Haiku, which is fast.

Yes, Fable tends to only shine when the work to be done is complex and it takes a long time. Other models wedge in different ways.

I do wish the subscription had a separate weekly allocation for rare usage.

Yeah, it’s not great—except for debugging. It shines there.

Keep composer away from anything configuration related—it will ruin your day.

We know this one is nudged by one persons twitter account.

This is, after all, SpaceTwitterAI.

Use console and pay API rates. Very predictable.

Much larger, much more expensive. It’s a new tier.