p1esk

7 ms·

91.
by p1esk 5mo ago
From TFA: “ChatGPT came up with an idea which is original and clever. It is the sort of idea I would be very proud to come up with after a week or two of pondering, and it took ChatGPT less than an hour to find and prove”
92.
by p1esk 5mo ago
Good music does not need to be complex
93.
by p1esk 5mo ago
I’ve been using llm as my personal pcp for 3 years now. I’m extremely pleased with the results.
94.
by p1esk 5mo ago
I cannot wait until doctors are fully automated. Shouldn’t be long now, hopefully just a few years.
95.
by p1esk 6mo ago
I fully expect to buy one within the next three years. Probably Optimus 4, depending on the price.
96.
by p1esk 6mo ago
2-3 years. Probably Optimus 4 will be able to do that, but probably not very well - kind of like tesla fsd drives currently.
97.
by p1esk 6mo ago
What new laws would you like to see?
98.
by p1esk 6mo ago
Depends on fast you want it to be. I’m guessing a couple of $10k mac studio boxes could run it, but probably not fast enough to enjoy using it.
99.
by p1esk 6mo ago
Do you think a lot of people have “systems” to run a 1.6T model?
100.
by p1esk 6mo ago
AI will seemingly always be ... You do realize that the whole image generation field is barely 10 years old? I remember how I was able to generate mnist digits for the first time about 10 years ago - that seemed almost like magic!
101.
by p1esk 6mo ago
how? From GP comment: “either start an open source project, or become the main collaborator in one.”
102.
by p1esk 6mo ago
This is true in many industries. Try to build a house without tools.
103.
by p1esk 6mo ago
The problem is - the board can’t fire him
104.
by p1esk 6mo ago
Someone from AMD posted this a few minutes ago, then deleted it: "Anush's success is due to opting out of internal bureaucracy than anything else. most Claude use at AMD goes through internal infrastructure that can take hundreds
105.
by p1esk 6mo ago
We are talking about running a business. In business world no one ever cared where code is coming from, the only concern is how much the code costs.
106.
by p1esk 6mo ago
your choices are to pay up or shut down. Another choice is to switch to a different model, perhaps open source this time.
107.
by p1esk 6mo ago
I’d stop asking it to update session logs
108.
by p1esk 6mo ago
Yep, can confirm - just today, when debugging a failing test, Opus on high effort in CC repeatedly made stupid moves, such as running a different test instead of the failing one, and declaring that the failure is non-deterministic and canno
109.
by p1esk 6mo ago
You might have to pay tax on that
110.
by p1esk 6mo ago
My policy is the opposite – I switch companies every two years. Especially if I get too comfortable. Usually because I’m bored or because the company grows too much (>100 people is too much for me).
111.
by p1esk 6mo ago
What’s wrong with reading HN in a browser? What problem are you trying to solve?
112.
by p1esk 6mo ago
Just to clarify - it’s a temporary effect - lasts for 3-6 months
113.
by p1esk 6mo ago
Oh ok, makes sense then
114.
by p1esk 6mo ago
Wow, this book has been published in 2025, and it has zero mention of AI generated music. Not saying it's a bad thing - from the table of content it covers a lot of important fundamentals, but ignoring the elephant in the room is... we
115.
by p1esk 6mo ago
It’s so ironic that Apple still publishes AI research and OpenAI does not.
116.
by p1esk 6mo ago
LLMs are orders of magnitude simpler than brains, and we literally designed them from scratch. Also, we have full control over their operation and we can trace every signal. Are you surprised we understand them better than brains?
117.
by p1esk 7mo ago
I’ve been a paid subscriber for all three players since day 1. CC (Opus) has been a clear winner for agentic coding starting about 6 months ago. GPT5.4 reduced the gap somewhat but the gap is still there.
118.
by p1esk 7mo ago
Yes, this is a problem - the most challenging samples might not even be present in your training data. This means your model will not perform well if real world data has lots of challenging samples. This can be partially solved if we make s
119.
by p1esk 7mo ago
I don’t see why it’s not ok to do that to an AI model. Or are you asking why they don’t want you to do it?
120.
by p1esk 7mo ago
Maybe I don’t understand this data labeling issue - are you talking about imbalanced classification dataset? Are hard classes under-represented or missing labels completely?

More ›