Head of AI at @clay • prev https://nitter.cf/t.co/LmhUuxxmrS (YC W21) @amazon @pennmandt
New York
Joined April 2009
- Tweets3.7K
- Following3.4K
- Followers2.7K
- Likes47.7K
Jeff Barg retweeted
What brutal feedback do you have for @clay and I’ll route to the right team
Jeff Barg retweeted
Introducing auto-mode for Claygent. It picks the AI model for each row of research, balancing answer quality AND cost.
Now on by default for new Claygent columns:
- 47% cheaper than sending every task to the most capable model.
- <1 second to choose the model for each research task.
- You can still pick your favorite model anytime.
Super cool thing:
We post-trained an open-weight model with @appliedcompute to handle the routing. It runs on their AC2 platform and handles hundreds of millions of queries a month.
@jeffbarg, Vyshnavi, @rhythmrg, and @_brylee10 break it down below.
Jeff Barg retweeted
A year ago, Goodlabs was in one blood center, one day a week.
Today we're in 19 centers across six states, open 7 days a week, with more than 70,000 people on the platform.
Thank you to our donors and partners for making this possible.
We're so back
can @typesafeai Jev count the r's in strawberry?
no.
47% says 3, 47% says 2. a coin flip, same as every LLM.
On 70% on 168 test words, it undercounts doubled letters like everyone
then I gave it the letters instead of the word: ["s","t","r","a","w","b","e","r","r","y"]
168/168. same model, same question, 260ms
@CompleteSkeptic is this expected?
“This is obviously a toy” famous last words
I've seen people describe Jev as an "AI if statement". But what if it actually WAS an if statement?
Introducing Probably: a programming language powered by Jev: probably-lang.southpolesteve…
Jev baked into the language. “feels” asks a question. “match” routes between descriptions. “while” keeps going until something stops feeling true.
This is obviously a toy, but it's fun to think about what something like Jev unlocks. Jev makes the decisions, an LLM does the writing, and a little program ties it together.
The iPhone Duo I want looks a lot more like two Apple Watch Ultras stitched together than an iPad mini folded in half
- Millennium prize is $1m
- 130b output tokens = $6.5m on Astra pricing
OpenAI really still is a non-profit
openai.com/index/navier-stok…
🐐
I’ve joined Anthropic.
In the space of possible tokens, there are surprising, interesting, and true things waiting to be discovered.
I’m excited to help make Claude think long and hard enough to discover them.
To my friends at humans&, it was incredible working with all of you, and I can’t wait to see what you build next.
To celebrate this Labor Day, I will be turning off Fast Mode in Codex to give Astra a much-needed break
Jeff Barg retweeted
Life update: started a job as a creator-in-residence at @clay_gtm.
I get to build, research, and storytell for a living😀
For my 1st YT video, I went down a silly rabbithole that turned quite meaningful.
I've never met a Bhaumik in my life, I thought Clay could fix that
Jeff Barg retweeted
In 13 minutes, @jeffbarg, Vyshu Khota, and Soroush Khadem walk through how Clay scaled agent evals agents at 300M+ runs a month.
Topics covered:
✅ Their four quadrant eval framework
✅ Why closing the production-to-eval loop is the hardest part
✅ How a data lake and long context changed what agents can do with data