joshuastuden

11 ms·

1.
by joshuastuden 1mo ago
Exactly. I saw them using things from 2025... AI sorta sucked then and didn't really "take off" until that Opus drop in December or whatever it was.
2.
by joshuastuden 2mo ago
That seems to be because Apple's AI division sucks. OpenAI came in 2018 and chatGPT 2.0 was already way better than anything Apple ever did.
3.
by joshuastuden 3mo ago
Agent-shell.el for the CLI experience that's integrated with emacs. You can send things from emacs to your Claude code prompt and it will include the file and line numbers. There's also claude-code-ide.el that gives you more of a
4.
by joshuastuden 3mo ago
Agent-shell.el and claude-code-ide.el Agent-shell is like the Claude code CLI. Claude-code-ide is like cursor with ghost completions/auto complete, chat, etc.
5.
by joshuastuden 3mo ago
Emacs can give you all the vim key bindings. Vim can't do that.
6.
by joshuastuden 3mo ago
Yes to some extent. When you attach to it, the default screen will still be scratch, but all your buffers are still open. There are some settings you can set to fix that, if it's burdensome.
7.
by joshuastuden 3mo ago
Emacs --daemon once and then attach to it with emacsclient.
8.
by joshuastuden 3mo ago
I don't understand what this means. How does claude-code-ide support or need a terminal like ghostel?
9.
by joshuastuden 3mo ago
Also, if a slash command is missing, that's because the model's ACP client doesn't implement it: https://github.com/xenodium/agent-shell#why-doesnt-agent-she...
10.
by joshuastuden 3mo ago
C-c C-v to change the model or M-x agent-shell-set-session-model
11.
by joshuastuden 3mo ago
When on the agent shell buffer, do C-c C-v with the latest agent-shell to change the model
12.
by joshuastuden 3mo ago
Anthropic scrapped their sdk billing thing
13.
by joshuastuden 3mo ago
Emacs evil-mode gives you vim bindings.
14.
by joshuastuden 3mo ago
Why not use agent-shell? It makes the whole experience great. Also claude-code-ide.el. try these
15.
by joshuastuden 3mo ago
GPT 5.5 regularly got things correct where opus 4.8 got things wrong. Just today, it was launching an executable and putting the environments after the launch. It should have been Env1=val env2=val ./executable .... But instead it did
16.
by joshuastuden 3mo ago
>The projections, which were reviewed by The Wall Street Journal No, they're just projections of profitability
17.
by joshuastuden 3mo ago
Did you read it? It just said AI assisted with the process. So did the Internet. AI didn't come up with the solution or even the idea. It simply helped prune the search space, which is valuable, but not worth the trillions of dollars i
18.
by joshuastuden 3mo ago
They already fold proteins with simulations. Just because people are using it doesn't mean it's a net good. Lots of people use social media and that's just rotting brains and making people far more polarized than they ever we
19.
by joshuastuden 3mo ago
They haven't done humanity a favor at all. The innovation that these LLMs have produced has been small. A few fun math theorems where the answer was gleaned from a pattern in the training data. Great,.but it doesn't change the wor
20.
by joshuastuden 3mo ago
They're not really losing ground to Anthropic. 5.5 was a bit better than 4.8. Fable was good, and was a jump over 4.8, but only incremental over 5.5. Anthropic is also likely losing money, right?
21.
by joshuastuden 3mo ago
SpaceX also wants to put data centers in space. That's the big market for SpaceX and how it ties into AI.
22.
by joshuastuden 3mo ago
Signs point to them not caring.
23.
by joshuastuden 3mo ago
They certainly can restrict technology that is deemed to be a national security risk.
24.
by joshuastuden 5mo ago
Semianalysis just argues for the industry case. They never go against the grain.
25.
by joshuastuden 5mo ago
The CTO is more creative than the engineer so who needs the engineer.
26.
by joshuastuden 5mo ago
I'm not asking to say "it's not because of LLMs" I am asking for evidence that LLMs are creating revenue for users.
27.
by joshuastuden 5mo ago
You're basically arguing for massive headcount reductions.
28.
by joshuastuden 5mo ago
Exactly. At Cerebras I know of several people who burn tokens on completely USELESS tasks (randomly changing pixels in an image) just to keep them high up on the token leaderboard. I suspect the other tokenboard leaders are doing the same.
29.
by joshuastuden 5mo ago
How can that be attributed to any code an LLM wrote? >$8 billion of net income was the result of a tax benefit the company realized in the first quarter of the year. So exactly how much of their revenue is because of any code LLMs wrote
30.
by joshuastuden 5mo ago
How do you know they didn't?

More ›