scottydog51834
scottydog51834 on Hacker News
Slightly off-topic, but I am curious how difficult it is in general to receive an entrepreneurial visa to the US? I chatted with a potential co-founder, here now for a masters, who would seek out this visa if we were to start a company together.
Some people (namely the EAs) care because they don't want AI to kill us.
Another reason is to understand how our models make important decisions. If we one day use models to help make medical diagnoses or loan decisions, we'd like to know why the decision was made to ensure accuracy and/or fairness.
Others care because understanding models could allow us to build better models.
There's a field called Interpretability (sometimes "Mechanistic Interpretability") which researches how weights inside of a neural network function. From what I can tell, Anthropic has the largest team working on this [0]. OpenAI has a small team inside their SuperAlignment org working on this. Alphabet has at least one team on this (not sure if this is Deepmind or Deepmind-Google or just Google). There are a handful of professors, PhD students, and independent researchers working on this (myself included); also, there are a few small labs working on this.
At least half of this interest overlaps with Effective Altruism's fears that AI could one day cause considerable harm to the human race. Some researchers and labs are funded by EA charities such as Long Term Future Fund and Open Philanthropy.
There is the occasional hackathon on Interpretability [1].
Here's an overview talk about it by one of the most-known researchers in the field [2].
[0] https://transformer-circuits.pub/2021/framework/index.html
[1] https://alignmentjam.com/jam/interpretability
[2] https://drive.google.com/file/d/1hwjAK3lWnDRBtbk3yLFL2DCK1Dg...
I was hoping that the title "Community Is the Future of AI" would refer to the exciting community that's growing around Generative AI in SF. There was an awesome hackathon this past weekend and a fun networking event on Friday. "Cerebral Valley" (not sure if this is a company or an unofficial org of people from hacker houses) is a major (but certainly not the only) source of such events [1]. VCs, Twitter influencers, early-stage founders, and college grads have all been organizing other events, and so far, it's been a magical community (and hopefully this community is a net positive for the future of AI).
[1] https://cerebralvalley.ai/
Switch teams or leave? I left my faang a few weeks ago and I'm glad I did (though I realize not everyone has this luxury). Can't say everything is figured out but I don't have to interact with my former manager or feel anger when leaders promise they care about slowing down but then assign more work. In terms of isolation, if you're in SF and looking for a friend, happy to grab coffee and talk about something that isn't tech.
In case anyone else is curious, this yt video helps me understand how Azure ML works and I think with this understanding, combined with the Github read me, it will be doable.
https://www.youtube.com/watch?v=yBVXR8G8Bg8
This is a really cool step but as someone without the suggested GPU, it isn't easy or one click for me yet.
I am hoping that someone makes a very simple Jupyter notebook where I can enter my RLHF file and select a few other settings and just run (on AWS or Azure; willing to pay per fine-tuned model say $100-$500 for cloud credits + notebook access).
I'd love a tool where I can upload a private dataset and RLHF a model (even better if the tool provides the pre-trained model) without me having to worry about GPUs, memory, commercial access, or even writing any python code. I'd happy pay several hundred and maybe several thousands for access to this.
Starting 1-2 months ago, many people, myself included, have started saying "SF is back" after observing the excitement surrounding, attendance at, and quantity of Generative AI events. It's too early to know how permanent this effect will be, but an article discussing SF collapse should at least mention the possibility that decline has ended and attempt to explain why event attendance won't be the first step towards recovery.