Omar Khattab
Right now, it's much easier to *build* a system for most NLP tasks than it is to *evaluate* a system for it.

And this is a major bottleneck. A lot more focus is needed on capabilities evaluation.

Maybe recurring TREC-style competitions and annotations campaigns in NLP?
tweet
Alan Tikwart Alden
My wife and I were just commenting this week that ⁦@iga_swiatek⁩ deserved an upgrade in the clothing department…
and ⁦@BenShelton⁩ is going to be a treasure to watch for years to come…well done ⁦@on_running

#tennis https://t.co/TKdm2ZLDxe
tweet
Jay Hack
⚡️Exciting ideas from tonight's "AI Tinkerers Meetup"

1) LLMs for dementia diagnosis/treatment

- Offer seniors conversational partners for company

- Use chat to diagnose dementia

- Perform personalized conversational therapy to keep mind agile
tweet
Offshore
Photo
AK
RT @Gradio: Notice anything different? 👀 https://t.co/WemkmtY7LL
tweet
Offshore
Photo
Adam.GPT
RT @araghougassian: I fed chatGPT some credit card transactional data to see if it could help me by acting like an analyst. Consider me blown away.. https://t.co/mnEbYmuu5u
tweet
Offshore
Photo
Stephen Young
RT @proximasan: @Microsoft designer is cool but the design suggestions are kinda cheugy https://t.co/yDi4wCDWeF
tweet
Offshore
Photo
Carlos E. Perez
RT @ruinwanderer: China changed the names of 9 Russian places for "historic names" in Chinese.

There's a directive to this effect signed on February 6, 2023.

The Russian @mfa_russia that usually suffers word diarrhea didn't even comment on it for fear of upsetting the Chinese.

So much for… https://t.co/h3cOlUabRJ https://t.co/NX4XspcSly
tweet
Carlos E. Perez
RT @IntuitMachine: 1/n Modularity is essential for any disruptive technology. For years deep learning lacked sufficient remix functionality to quickly customize solutions. Everything had to either be trained from scratch or fine-tuned. The latest innovations are tearing down these restrictions.
tweet
Sam Altman
We offer no explanation as to why [anything works except] divine benevolence.
tweet