technics256

8 ms·

31.
by technics256 3y ago
This also jumped out at me. It also seems that they are selectively choosing different promoting strategies too, one lists "CoT@32". Makes it seem like they really needed to get creative to have it beat GPT4. Not a good sign imho
32.
by technics256 3y ago
I came here to check why my merge status wasn't updating and yep, HN breaks news the fastest still. With all of the GithHub Universe AI announcements, it's sad to see the main github product being ignored.
33.
by technics256 3y ago
It's listed in the graph titled "hard to answer"
34.
Claude 2.1(anthropic.com)
525 points by technics256 3y ago|322 comments
35.
by technics256 3y ago
At the hashi conf last month they started talking about adding terraform features that would only work on terraform cloud. It was very sad to see
36.
by technics256 3y ago
Is there any movement in the language spec to address this in the future with Nil types or something?
37.
by technics256 3y ago
I even tried it myself, lying in bed and creating sexual fantasies with Dolores through text, culminating in ejaculation. This was an excellent read. And I applaud the creator for experimenting with this to the fullest.
38.
by technics256 3y ago
Any mirrors we can use for images like golang, etc?
39.
by technics256 3y ago
Off topic but in general how does something like this compare to cloud hosted ocr solutions?
40.
by technics256 3y ago
This was a very touching piece. Thanks for sharing.
41.
by technics256 3y ago
Haha yes. We even found a magic eye/autostereogram of goatse once. Was pretty epic. "Hey come check out this amazing magic eye". Boom!
42.
by technics256 3y ago
I saw it the first time as a kid. My friends and I tried to trick each other to see it too. I'm fine. We're fine.
43.
by technics256 3y ago
If you have a mac, Cleanshot X I find is better across the board cleanshot.com
44.
by technics256 3y ago
Is there a guide for integrating this in local dev, either locally or if you want to view it on the hosted? Ideally hosted, devs can bring up our app locally, and view their logs and traces etc when testing and building
45.
by technics256 3y ago
go away ai bot
46.
by technics256 3y ago
That's funny, I've only ever used Atlantis with a smattering of tfc
47.
by technics256 3y ago
This gave me very happy proustian moments
48.
by technics256 3y ago
In my experience it is very hollow. It skips details unless you force it to. Gpt4 is still way better
49.
Claude 2(anthropic.com)
433 points by technics256 3y ago|260 comments
50.
by technics256 3y ago
Guess it got brute forced
51.
by technics256 3y ago
I really liked phind, but the new model doesn't compare to GPT4. I'd gladly, gladly pay to get the original phind back.
52.
by technics256 3y ago
it's also really, really bad and fails compared to even open source models right now.
53.
by technics256 3y ago
Can you elaborate? Just started building out a new platform centered on it
54.
by technics256 3y ago
Do you have any example papers i can play with?
55.
by technics256 3y ago
They can, and they are trying: https://www.washingtonpost.com/health/2023/04/10/operation-w... https://erictopol.substack.com/p/project-next-gen-the-united...
56.
by technics256 3y ago
You likely got rejected. Was the same for me and I reapplied with a good use case and was let in
57.
by technics256 3y ago
I have access to claude. It's not bad, but decently behind gpt4 for code
58.
by technics256 3y ago
PaLM 2 on HumanEval coding benchmark (0 shot): 37.6% success GPT-4: 67% success Not even close, gpt4 miles ahead
59.
by technics256 3y ago
it did? not for me on my admin workspace account--how long did it take? :/
60.
by technics256 3y ago
Curious if you tried Vespa?

More ›