uptownfunk

7 ms·

91.
by uptownfunk 2y ago
It’s marketing to get the best researchers. The researchers want the meta pay and they want to hedge their careers to continue to publish. That’s the real game, it’s a war for talent. Everything else is just secondary effects.
92.
by uptownfunk 2y ago
It’s for VCs to go ooh and aah and boost their valuation.
93.
by uptownfunk 2y ago
This hits the nail on the head. It is a consumer facing product not a technology to solve deep thinking.
94.
by uptownfunk 2y ago
Thanks for your work and also for your comments of AF3 and Chai-1. It sounds like you are implying there are potentially gross and subtle types of data set leakages taking place between the train and test which are resulting in what seem to
95.
by uptownfunk 2y ago
What happens when the new models come out and the data centers are full of old models to be decommissioned. Would love to buy a huge amount of h200 once they have become “obsolete”
96.
by uptownfunk 2y ago
Wow the end of an era. I first heard about it at Cal through college roommates. I haven’t used it much lately but still.. sad to see it go.
97.
by uptownfunk 2y ago
A little bump of cbd or thc 5-10mg and the induced sleep, it’s a good reset.
98.
by uptownfunk 2y ago
Like if it really worked and made my wife go wow that’s a good photo, I would pay for it, probably 1.99-2.99 a month, or maybe 19.99 one time.
99.
by uptownfunk 2y ago
I always suck. Like if an app could just give me a green check when the photo in frame was good and then I could just click it then that’d be great.
100.
by uptownfunk 2y ago
Is it possible to train an LLM that is minimally biased and that could assume various personas for the purpose of the experiments? Then I imagine it’s just some prompt engineering no?
101.
by uptownfunk 2y ago
Not the author but thought it looked interesting and wanted to share. Has anyone built something similar?
102.
GitGPT A natural language command line Git assistant(github.com)
2 points by uptownfunk 2y ago|1 comments
103.
by uptownfunk 2y ago
Are they that bored over there at amazon.
104.
by uptownfunk 2y ago
The correct solution is to have IT force push updates only when they deem fit (after they have tested internally on some ghost machines).
105.
by uptownfunk 2y ago
It’s really really bad out there. If you have a job be thankful. If you don’t, I wish you strength and support.
106.
by uptownfunk 2y ago
Whatever anyone will say. I am glad he is doing something meaningful with his life beyond just making investors in OpenAI more wealthy. At some point I suspect the money just becomes a number and then you can afford to think about bigger th
107.
by uptownfunk 2y ago
Very clearly the business motive make sense, go after nvidia gpu monopoly. Can someone help a lay person understand the pitfalls here that prevent this from being an intelligent venture?
108.
by uptownfunk 2y ago
Smells fishy anti trust
109.
by uptownfunk 2y ago
I just want a Python version of rstudio- I bet they probably support Python by now
110.
by uptownfunk 2y ago
why does everyone make their own note taking app - is it easier to specify the app in code than to build an app that gives you what you want?
111.
Ask HN: What RAG setup gets me 95% of the way there?
6 points by uptownfunk 2y ago|5 comments
112.
Ask HN: Anyone here burnt out of the AI hype?
25 points by uptownfunk 2y ago|43 comments
113.
by uptownfunk 2y ago
Agree with the other commenter the new approaches are mostly neurosymbolic and based on DNNs vs the approaches given here. FWIW I think this is one of the critical pieces to AGI and I know many of the large AI labs are working on integratin
114.
by uptownfunk 2y ago
Remember folks there is no free lunch :)
115.
by uptownfunk 2y ago
Entrepreneurship is so brutal, so crushing. Day job way way easier.
116.
by uptownfunk 2y ago
Very helpful, so the type of work doesn’t really matter ?
117.
by uptownfunk 2y ago
Ha ha at “delving” common gpt language
118.
Ask HN: To biz people here: what problems does your eng team push back on?
3 points by uptownfunk 2y ago|4 comments
119.
by uptownfunk 2y ago
I am more interested in meta-cognition, advanced cognitive architectures, learning about how humans learn and then figuring out how we get machines to do that better and faster
120.
by uptownfunk 2y ago
Arc agi is a small stepping stone to agi but is not agi. Program search mimics what humans do to a certain extent but not in entirety. A more general world model and reference will be required for agi.

More ›