Lightnews — Scholar-powered news

Eugene Vinitsky 🍒

@eugenevinitsky.bsky.social

10K followers 2.7K following 5.1K posts

Anti-cynic. Towards a weirder future. Reinforcement Learning, Autonomous Vehicles, transportation systems, the works. Asst. Prof at NYU https://emerge-lab.github.io https://www.admonymous.co/eugenevinitsky

emerge-lab.github.io

Posts Media Videos Starter Packs

Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 22m

I think something close to what Shubhendu says here: bsky.app/profile/shub...

It’s not a bubble in that people are correct about the value of the tool

Shubhendu Trivedi @shubhendu.bsky.social · 32m

I think it's not a bubble in the sense of the south sea bubble, or the Dutch tulip mania. Those were actual bubbles--where "actual" means that they were almost fully driven by fraud. In contrast, the bubbly aspect of AI comes from the fact that a huge volume of risk capital is all in one space.

Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 1h

One of the ways you can tell I’ve been in ML for slightly too long is the usage of the word “humans“ as a category

1 3

Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 1h

It’s important to understand this! It’s either automation or no social media at all. It’s fine to make the latter choice but there aren’t really alternatives

1 3

Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 1h

The internet is unfortunately a flood of horrific content that is filtered out of your view. When twitter trust and safety ceased to exist for a few weeks, the site was instantly flooded with horrible stuff

1 3

Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 1h

For image moderation systems there are only three possible choices:
1: automated / “AI”
2: make humans see horrible things
3: no images at all

2 3 26

Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 1h

I don’t even think it’s a bubble in any meaningful sense but yes please give me half price H100s I will do so much good stuff with them

4 19

Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 1h

Do you know how happy I would be if the AI ”bubble” collapsed? My god yes give me cheap GPUs to snatch up

6 6 91

Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 1h

This site generates a very particular type of bait. Total nonsense said by someone who you’d think could be convinced otherwise.

Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 9h

I think a lot of this stuff is nearly there, someone else could build it for them at this point

Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 9h

there are still so many things to try

Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 9h

Soooooo when is bluesky deploying bridging systems
arxiv.org/abs/2301.09976

Bridging Systems: Open Problems for Countering Destructive Divisiveness across Ranking, Recommenders, and Governance

Divisiveness appears to be increasing in much of the world, leading to concern about political violence and a decreasing capacity to collaboratively address large-scale societal challenges. In this wo...

arxiv.org

1 2 10

Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 9h

oops, meant to write bridging systems

Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 9h

Deleted the post because it was meaner than I'd like to be. Just a temporary fit of anger

Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 12h

Hm I’m wary of giving advice that probably should come from a professional. For me personally the signs are usually akin to depression I.e I don’t want to do anything at all

1 3

Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 12h

This is a joke I hate this

Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 12h

7 minute drive 2 hour walk that’s American efficiency for you

1 4

Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 13h

Yes, while it is not possible to detect LLM text in general it is very obvious when a message is in ChatGPT voice

Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 13h

Reading LLM emails just makes me really sad. I want to know what the person’s voice actually sounds like, not the machine’s

1 22

Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 16h

I burned out pretty hard towards the end of grad school and it was a bit of a road back

Reposted by Eugene Vinitsky 🍒

Ramon Astudillo @ramon-astudillo.bsky.social · 17h

(repost welcome) The Generative Model Alignment team at IBM Research is looking for next summer interns! Two candidates for two topics

🍰Reinforcement Learning environments for LLMs

🐎Speculative and non-auto regressive generation for LLMs

interested/curious? DM or email [email protected]

1 10 13

Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 19h

Good on you. I wasn’t really speaking of making money, just trying to figure out how to interpret a prediction as more than hot air

2 2

Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 19h

My personal experience of it is that it’s quite challenging for me to push myself as hard as I used to because anxiety kicks in

2 22

Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 19h

Besides stocks though, what else? Would you take a job that pays the maximum possible salary or the opposite because money would be worthless? I don’t know how to draw out what such a person should believe

2 3

Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 19h

I sort of view it as a proxy for most white collar work which is why I think the question is interesting. What should I expect of someone who is claiming white collar work will be automated away

2 1

Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 19h

If you’re staring burnout in the face, it’s a great moment to step back a bit

1 20