thomashop

5 ms·

1.
by thomashop 2mo ago
I'm also a fan of the Yamanote Line. I made a psychedelic AI audio-visual collage inspired by it. https://www.youtube.com/watch?v=xwUSzUvShqcaa I made field recordings during my last stay in Tokyo. From those, I made a
2.
by thomashop 4mo ago
i thought the agent can execute real ffmpeg to compare
3.
Shprout – 23-line coding agent in bash. The script is its own prompt(github.com)
4 points by thomashop 5mo ago|0 comments
4.
by thomashop 6mo ago
I don't think it's correct that we destroyed everything that isn't us. If we take all living beings, we have destroyed only a small percentage.
5.
by thomashop 10mo ago
Ah understood i misread
6.
by thomashop 10mo ago
Why is privacy a concern? You can run them in your own infrastructure
7.
by thomashop 1y ago
I'm also a fan of the Yamanote Line. I made a psychedelic AI audio-visual collage inspired by it. https://www.youtube.com/watch?v=xwUSzUvShqcaa I made field recordings during my last stay in Tokyo. From those, I made a
8.
by thomashop 1y ago
I have done quite a bit of game dev with LLMs and have very rarely run into the problem you mention. I've been surprised by how easily LLMs will create even harmful narratives if I ask them to code them as a game.
9.
by thomashop 1y ago
The rest was fine without citations, but the part you disagree with needs them?
10.
by thomashop 1y ago
I have the impression that the thinking helps even if the actual content of the thinking output is nonsense. It awards more cycles to the model to think about the problem.
11.
by thomashop 1y ago
I use AI to simplify code. My manifesto has always been code is debt. Works really well too.
12.
by thomashop 1y ago
I don't care that much. This level of knowledge about me can also be easily found on the internet. I'm also working almost entirely on open-source software so I'm happy if the AIs know more about my projects. But this, of cou
13.
by thomashop 1y ago
Interesting { "assistant_response_preferences": { "1": "User prefers concise responses for direct factual queries but detailed, iterative explanations when exploring complex topics. They of
14.
by thomashop 1y ago
Who runs python code on mobile?
15.
by thomashop 1y ago
The AI is not paying for watching videos yet
16.
by thomashop 1y ago
But aren't they getting this data already at a much larger scale? GPT is still one of the backbones in many coding assistants, even Windsurf.
17.
by thomashop 1y ago
I'm pretty sure the model is cached with the system prompt already processed. So you should only pay extra tokens.
18.
by thomashop 1y ago
"I cannot emphasize this enough: ChatGPT is not generating meaning. It is arranging word patterns. I could tell GPT to add in an anomaly for the 1970s - like the girl looking at Billy’s Instagram - and it would introduce it into the te
19.
by thomashop 2y ago
It always feels like people are irrationally critical of AI assisted stuff. Does the typical Hacker News comment have more substance? - Informally benchmarked against 4 specific competitors: Gemini, OpenAI, o3, and Claude - Identified two c
20.
by thomashop 2y ago
Why sorry? So what? I often write things I want to post in bullets and then have it formulated better than I could by an LLM. But its just applying a style. The content comes from me. My wife is dyslexic so she passes most things she writes
21.
by thomashop 2y ago
It feels like we've been moving in the opposite direction, where more and more models from various countries are state-of-the-art. The idea that there will be one model to rule them seems very unlikely.
22.
by thomashop 2y ago
Sure. You can always write more tests. That's not a problem specific to AI. I'd also do code reviews on the code AI produces.
23.
by thomashop 2y ago
Why not write tests with AI, too? Since using LLMs as coding assistants, my codebases have much more thorough documentation, testing and code coverage. Don't start when you're already in a buggy dead-end. Test-driven development w
24.
by thomashop 2y ago
> And that's just hooey. Well, I mean Google animism. It's hooey from your point of view.
25.
by thomashop 2y ago
I'm one of those people. Very happy to associate myself with animism and anthropomorphize animals and machines. I think one of the biggest mistakes with Christianity and the western world is that we see ourselves as something greater t
26.
by thomashop 2y ago
Ask it to plot the graph with python plotting utilities. Not using its image generator. I think you need a ChatGPT subscription though for it to be able to run python code.
27.
by thomashop 2y ago
I made a similar experiment using the Hitchhiker's Guide to the Galaxy theme. Have you ever argued with a neurotic elevator? Try convincing this Sirius Cybernetics Happy Vertical People Transporter to go down! There are many ways to pe
28.
by thomashop 2y ago
New research shows that by extending instruction tuning to handle visual tokens, LLMs can simultaneously learn image understanding and generation with minimal changes. The most intriguing finding is that visual generation capabilities emerg
29.
MetaMorph – Language Models Are Closer to Being Universal Models Than We Thought(tsb0601.github.io)
1 points by thomashop 2y ago|1 comments
30.
by thomashop 2y ago
We have moved 2 orders of magnitude in the last year. Not that unreasonable

More ›