Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
tough
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
29 ms
·
841.
▲
by
tough
2y ago
three.js then? godot seems cool, would love to get some vision lite dev kit in a few years to play with
842.
▲
by
tough
2y ago
arent all apps just expo react webui wrappers nowadays?
843.
▲
by
tough
2y ago
Confirm it works with mac! gtk-chat at least, having some issues with the notif lib for gtk-applet screenshot: https://postimg.cc/KKxQNdG6
844.
▲
by
tough
2y ago
does the postinstall script step has anything to do with this? i noticed bun doesn't run them by default unless you whitelist them
845.
▲
by
tough
2y ago
http://nfx.com/ was always a good source on Network Effects on Digital Products. I always also liked Andrew's Chen blog posts, he ended up hired by a16z
846.
▲
by
tough
2y ago
Also PayPal famously started doing funny things forcing sellers to use it once it bought out Ebay. But yeah you have always to have in mind who is writing what you read and what the unsaid things are, great points
847.
▲
by
tough
2y ago
They have also been giving back to postgres some of their extra work, and also their real time stuff i think is on erlang? I agree you might prefer to choose the stack yourself, but for total n00bs and vibe coders supabase is a great start
848.
▲
by
tough
2y ago
AGPL tries to solve this
849.
▲
by
tough
2y ago
need it here for cli usage! https://github.com/agustif/llm-lmstudio
850.
▲
by
tough
2y ago
or by replying you expose yourself to handing -proof- of the origins of the training data set to the copyright owner wanting to sue you next
851.
▲
by
tough
2y ago
Google's Gemma new QAT training comes from google using their own TPU's don't know much about the rest tho
852.
▲
by
tough
2y ago
lack of knowledge allows one to avoid what produces the suffering or depression for others to acquire such knowledge or relaizations by virtue of being dumb one protects itself from much of the fuck ups of the world by staying oblivious to
853.
▲
by
tough
2y ago
the point is when there’s a system prompt you didnt write you get autocomplete of your input + said dystem prompt, and as such biasing all outputs
854.
▲
by
tough
2y ago
Yea but in america such a transplant probably costs 300k just to go to the hospital ez prob also raising a child way expensier if you factor uni and such into it vs UK
855.
▲
by
tough
2y ago
tighter feedback loops the closer your shotting instances are to each other
856.
▲
by
tough
2y ago
fully-autonomous makes more sense in the agentic vicab imho at the end its fine if the agent self corrects amongst many shots too
857.
▲
by
tough
2y ago
hn auto shorting of titles makes funny voincidences some times totally rewriting the meaning of submissions titles by omission of just a few words heh
858.
▲
by
tough
2y ago
OpenAI launched codex 2 days ago, there's open forks already that support other providers too there's also claude code proxy's to run it on local llm's you can just do things
859.
▲
by
tough
2y ago
Open Codex (A codex fork) that supports gemini and openrouter providers https://github.com/ymichael/open-codex google models on cli are great.
860.
▲
by
tough
2y ago
didn't occur to me that was possible in the US cause private unis spend most on marketing i guess but glad it is stand corrected
861.
▲
Beyond Public Access in LLM Pre-Training Data: Book Content in OpenAI's Models [pdf]
(ssrc-static.s3.us-east-1.amazonaws.com)
1 points
by
tough
2y ago
|
0 comments
862.
▲
by
tough
2y ago
Research team from orielly found out openai trained on copyirghted books prob got a sub... https://ssrc-static.s3.us-east-1.amazonaws.com/OpenAI-Traini...
863.
▲
by
tough
2y ago
its not a nonprofit (cornell, its a university) but they do accept donations https://info.arxiv.org/about/donate.html
864.
▲
by
tough
2y ago
this is why you use row level security and proper authorization on postgres and not elastic search eith some email id as filter in 2025
865.
▲
by
tough
2y ago
they already started blocking some MSFT extensions on cursor saw another day a thread on hn about it
866.
▲
by
tough
2y ago
makes sense thanks for the feedback
867.
▲
by
tough
2y ago
not op but LLM as Judge is a thing https://arxiv.org/abs/2411.15594
868.
▲
by
tough
2y ago
Thanks, this looks great as i was hitting o3 api request limits making the -regular- codex unusable for my use-case PS: This fork aims to support other model providers
869.
▲
by
tough
2y ago
Im not mocking AI, and while the internet and smartphones fundamentally changed how societies operate, and AI will probably do so to, why the Doomerism? Isn't that how tech works? We invent new tech and use it and so on? What makes AI
870.
▲
by
tough
2y ago
I can see how students would be first users of this kinda of tech but am not on those spheres, but I believe you. As per spammy applications, hasn't always been this the case and now made worse due to the cheapness of -generating- plau
More ›