practice9

4 ms·

1.
by practice9 12d ago
You rather generously assume engineers are in touch with their systems. Even before layoffs many teams just maintained things org has long lost coherent knowledge of After layoffs and typical org knowledge churn - you can either rewrite i
2.
by practice9 4mo ago
It's because HN is in AI meta-psychosis :) Our experience is very similar except we didn't really have a review process before, and now LLMs find bugs before PRs get merged in main. We had 5x-100x speedups in some legacy but impor
3.
by practice9 11mo ago
LLMs are getting quite good at reviewing the results and implementations, though
4.
by practice9 1y ago
I find it hilarious/sad that the 0.5x cheaper Ergo M575 has much better design in that regard (just plastic that doesn’t degrade)
5.
by practice9 1y ago
They should have used Claude Code for reviews
6.
by practice9 1y ago
Humans cannot reason about code at scale. Unless you add scaffolding like diagrams and maps and … Things that most teams don’t do or half-ass
7.
by practice9 1y ago
A variation of “no taxation without representation”?
8.
by practice9 1y ago
The human is a bad co-author here really. I deployed lots of high performance, clean, well documented etc code generated by Claude or o3. I reviewed it wrt requirements, added tests and so on. Even with that in mind it allowed me to work 3x
9.
by practice9 1y ago
Well the system prompt is still the same for both models, right? Kinda points to people at OpenAI using o1/o3/o4 almost exclusively. That's why nobody noticed how cringe 4o has become
10.
by practice9 1y ago
Kinda similar in a way to China or Russia “disappearances”
11.
by practice9 2y ago
But who is the target group? Last time only some groups of enthusiasts were willing to work through bugs to even run the buggy release of Gemma Surely nobody runs this in production
12.
by practice9 2y ago
I tried the square example from the paper mentioned with o1-pro and it had no problem counting 4 nested squares… And the 5 square variation as well. So perhaps it is just a question of how much compute you are willing to throw at it
13.
by practice9 2y ago
With LLMs that even might be automated
14.
by practice9 2y ago
Well none of the labs have good frontend or mobile engineers or even infra engineers Anthropic is ahead in this because they keep their UIs simplistic so the failure modes are also simple (bad connection) OpenAI is just pushing half baked s
15.
by practice9 2y ago
One of those guys needs to be fined for the pump & dump scheme (with SPCE: Virgin Galactic), and the other one should be investigated if he was receiving money from the Russian government or influence agents.
16.
by practice9 2y ago
More like reparations
17.
by practice9 2y ago
The interesting thing is that ship was damaged almost immediately after leaving the port, had a chance to stop in Russian ports along the way but instead is doing a tour near EU countries. The crew is either amazingly incompetent or malicio
18.
by practice9 2y ago
It is cringe overenthusiastic, but a proper instructions/system prompt will fix that mostly
19.
by practice9 2y ago
I always double-check even the most obscure facts returned by GPT-4 and have yet to see a hallucination (as opposed to Claude Opus that sometimes made up historical facts). I doubt stuff interesting to kids would be so out of the data distr
20.
by practice9 2y ago
Wasn't it pointing to https://x.ai just a few months ago? Interesting
21.
by practice9 2y ago
teknium / Nous released Mistral finetunes (Hermes) that are quite great, and even published the datasets used for training. But for the worldsim I think they are really using Claude (probably Haiku or Sonnet) via openrouter ( https:&#x
22.
by practice9 3y ago
Interesting. I had a problem a few months ago with DNS not resolving Meta servers on my Starlink internet connection, but I was able to use the UI and the apps nonetheless, just couldn't open the store or update firmware. Seems like th
23.
by practice9 3y ago
Unless it's a recent change, it works perfectly fine offline (wifi turned off). As for alternatives, there is Pico, but Quest 3 may be superior in games selection. Or go wired which is of course less portable
24.
Documenting Devastation and Loss in Mariupol(hrw.org)
1 points by practice9 3y ago|0 comments
25.
by practice9 3y ago
Here is an interesting and relevant context: there is a huge amount of evidence that Roscosmos is taking an active part in the war effort. There is a great source about it from Eric Berger of Ars Technica: https://arstechnica.com
26.
by practice9 3y ago
But you can distill a wall of PR fluff to a concise message with ChatGPT... The problem is the average user thinks "long text = good" and their bosses often agree.
27.
by practice9 3y ago
Aren't most websites like Wirecutter in the business of paid reviews / affiliate marketing?
28.
by practice9 3y ago
That car's exterior is uglier than Cybertruck
29.
by practice9 3y ago
Wait, but the new title doesn't seem to be correct
30.
by practice9 3y ago
Yeah the Medprompt name is misleading

More ›