Technically impressive and fun to use. But as an avid music fan, I wouldn’t listen to these songs, since they’re just not interesting. I’m sure it’ll be useful for background music, but not anything I’d actively seek out.

I have a similar reaction to image generation: It’s more

Congrats to my sister!

I don’t know if OAI’s motivation is to create a “regulatory moat”, but it would be the likely outcome.

There are agencies (eg FTC) that regulate the bad business practices that AI will accelerate. The issue is lack of resources for serious enforcement.

nyti.ms/435J9Qt

Excellent summary around the dynamics of ranking by engagement.

Also Shubhanshu Mishra is looking for roles and I’d snap him up immediately. He’s a talented and humble research who isn’t afraid to roll up his sleeve and dive into tough engineering challenges. Also makes great math YouTube videos

For past 1.5 years we worked on building TweetNERD a dataset of 340K+ Tweets with Named Entities annotated and linked to Wikidata.
Join our poster at NeurIPS Conference in afternoon today.
Work with Aman Saini Raheleh Makki SM Ali Mollahosseini 🫡 Aria Haghighi
#NeurIPS22 #NLProc 🧵👇

This sucks for h1bs or others that can’t leave for reasons

Beyond that, this is good for others. Elon gets team who want or have to stick around and end the tweep vs goon dynamic. Others get incentive to move on. Not optimistic about product direction, but this beats status-quo.

SF-based, ex-Twitter Cortex Applied Research ML engineer looking for my next role. Love building products, connecting Eng with Research, working with colleagues. Open to full-time or contract. linkedin.com/in/kovas-bogut…

This bit on @jason is 🤌:“Last week, Mr. Musk dispatched a lieutenant to the ‘war room’ to ask Mr. Calacanis, who was there, to cool it on Twitter and stop acting as if he were leading product development or policy, people familiar with the exchange said.”
nytimes.com/2022/11/11/tec…

Dan Collier, Ph.D. Thanks a lot Dan. Means a lot at this time.

If you’re looking for an excellent ML EM who is available. Ali is your guy. He’s very strong technically and a great people manager.

Tweeps: my DMs are open. Please reach out if you need *anything*.

If you're hiring, I know a lot of great ML folks who are looking.

I don’t think there is a prettier song than Joni Mitchell’s “A Case of You”
youtu.be/0YuaZcylk_o

The GPT-3 prompt must have been “This tweet appeared in Stats 101 slides for a decade:”

New paper (w Aria Haghighi) on using Translation Pair Prediction to improve 0-Shot perf. from EN to non-latin langs e.g. HI, JA, AR at #WNUT2021 , #EMNLP2021 . Improves NER & Sentiment on tweets. We introduce a new translation source via Wikidata
Vid: youtube.com/watch?v=VR9qke… 1/n

Happy to share our EMNLP 2024 paper on incorporating social context in NLP models. Our approach merges graph embeddings with Transformer style language models to help the model disambiguate social context across time and geography.
Thread:👇🧵1/n
Code: github.com/twitter-resear…

How can you incorporate social factors (for eg. time, geography) which influence language use and understanding into large-scale LM's? With Shubhanshu Mishra and Aria Haghighi, we propose a simple pre-training method for this. arxiv.org/abs/2110.10319 (Findings of EMNLP 2021) #emnlp2021