dcre

18 ms·

331.
by dcre 1y ago
Change the url from x.com to xcancel.com to see it all.
332.
by dcre 1y ago
In my experience there is a real difference between the places available to rent and to buy. It is not an idealized single good with a spectrum of prices that line up cleanly with quality.
333.
by dcre 1y ago
Some other libraries I’ve been enjoying building CLIs with in TS that do more or less the same thing, though perhaps with slightly worse composability than Optique: https://cliffy.io/ https://github.com/tj&#
334.
by dcre 1y ago
Yes. Nearly everything they said they canceled either didn’t happen at all or the contract had ended before they canceled it.
335.
by dcre 1y ago
"Through July, DOGE said it has saved taxpayers $52.8 billion by canceling contracts, but of the $32.7 billion in actual claimed contract savings that POLITICO could verify, DOGE’s savings over that period were closer to $1.4 billion.&
336.
by dcre 1y ago
Wow!
337.
by dcre 1y ago
They’re talking about model quality. I still think they’re wrong, but the revenue is only indirectly relevant.
338.
by dcre 1y ago
Exactly.
339.
by dcre 1y ago
Value for the money is not at issue in the paper!
340.
by dcre 1y ago
Glad to see this out so quickly. Like I said[0] on the Gemini announcement post, it feels like Zed is trying to get out of the business of iterating on agent logic and just let other people handle it. Any prompting secret sauce a) is trivia
341.
by dcre 1y ago
Always fun to see a theoretical argument that something clearly already happening is impossible.
342.
by dcre 1y ago
It’s about 4 years old, twice as old as Zed!
343.
Is the "cost of inference" going up or down?(crespo.business)
9 points by dcre 1y ago|0 comments
344.
by dcre 1y ago
When you run Claude Code it lets you pick between API key and Claude subscription login. It sounds like it found an API key in your env and is trying to use that. You may have to find a way to trigger a login to get you into your subscripti
345.
by dcre 1y ago
I've paid a few dollars a month for my API usage for about 6 months.
346.
by dcre 1y ago
jj’s design is not constrained by git. git is one possible backend, but you don’t even need to use jj with git. I will eat a shoe if you can name something jj is prevented from doing by its relation to git.
347.
by dcre 1y ago
That is the only thing they seem to care about. It’s juvenile.
348.
by dcre 1y ago
My approach in "LLM-only RAG for small corpora" [0] was to mechanically make an outline version of all the documents _without_ an LLM, feed that to an LLM with the prompt to tell which docs are likely relevant, and then feed the e
349.
by dcre 1y ago
It is not about the present and future value of AI at all. It is about the present and future value of things other than AI. Here is the full quote: "There is nothing else after generative AI. There are no other hypergrowth markets lef
350.
by dcre 1y ago
Already in your first point you are mixing up two claims Ed also likes to mix up. The funny thing is these claims are in direct conflict with each other. There is the question of whether people find AI worth paying for given what they get.
351.
by dcre 1y ago
Are you referring to the post where I listed 4 claims and marked one ridiculous, one wrong, one unlikely, and one plausible? He is not wrong about everything. For example, after Sam Altman said in January that OpenAI would introduce a model
352.
by dcre 1y ago
Zitron is not a serious analyst. https://bsky.app/profile/davidcrespo.bsky.social/post/3lxale... https://bsky.app/profile/davidcrespo.bsky.social/post/3lo22k... https:/&#
353.
by dcre 1y ago
You're saying nothing dramatic or surprising has ever happened in history, or at best that when it did, no one saw it coming. Both are obviously wrong. In any case, Tooze is a serious thinker who has written various things besides this
354.
by dcre 1y ago
I take them to be saying the same thing — the difference is that Altman is referring to the training of the next model happening now, while Amodei is referring to the training months ago of the model you're currently earning money back
355.
by dcre 1y ago
That's interesting, though you have to imagine the data set is very low quality on average and distilling high quality training pairs out of it is very costly.
356.
by dcre 1y ago
I think the idea here is that gpt-5-mini is the cheap gpt-4 quality model they want to serve and make money on.
357.
by dcre 1y ago
I think this is debatable as more models become good enough for more tasks. Maybe a smaller proportion of tasks will require SOTA models. On the other hand, the set of tasks people want to use LLMs for will expand along with the capabilitie
358.
by dcre 1y ago
I don't understand why the absolute time period matters — all that matters is that you get enough time making money on inference to make up for the cost of training.
359.
by dcre 1y ago
I need to find the quote, but historian and political commentary Adam Tooze suggested a few months ago that China's global economic and political dominance in this century could be so total that future historians will come to think of
360.
by dcre 1y ago
Good point, very possible that Altman is excluding free tier as a marketing cost even if it loses more than they make on paid customers. On the other hand they may be able to cut free tier costs a lot by having the model router send queries

More ›