Adam Pollack (@Adam_Pollack_) header
Adam Pollack (@Adam_Pollack_) profile picture
Longmont, CO
Joined January 2020
Photo by Adam Pollack (@Adam_Pollack_)Photo by Adam Pollack (@Adam_Pollack_)Photo by Adam Pollack (@Adam_Pollack_)Photo by Adam Pollack (@Adam_Pollack_)Photo by Adam Pollack (@Adam_Pollack_)
In about 4 hours we’re going to get 20 posts saying how Fable 5.2 did something awesome but they only sent 2 messages and can’t chat with it more for another 2 days because of limits.
13
People love to point out how much market competition pushes frontier labs to out do one another driving price reductions and innovations, but this week the game theory is feeling much more like an oligopoly.
5
Adam Pollack retweeted
Labs; “we will have a nation of geniuses in a data center by 2028, capable of beating humans at every task, but you will need to hire our newly trained forward-deployed engineers for a six month engagement to deploy them to a single project”
17
13
2
230
24,011
Adam Pollack retweeted
to be clear, NVIDIA is NOT a car
88
193
86
2,663
289,373
5.4 is great but it has one glaringly bad quality that bothers everyone. Do you want to know what it is?
32
Vertical tabs on the Atlas app.
20
Leveraging my newborn’s first coos & smiles as paywall: family can watch the video after they read The Scaling Era: An Oral History of AI (2019–2025) by @dwarkesh_sp. They’ve dodged my AI reading list long enough. 🍼🤖
1
37
Adam Pollack retweeted
Replying to @bensrubin
this will have no unintended consequences! I love the future!
1
2
51
i love pete buttigieg and capitalism and LLMs and human artists and the trans community and my country and free speech and YIMBYism and conservationism and tradition and billionaires and globalism and american exceptionalism
56
20
5
749
51,495
Adam Pollack retweeted
Surround yourself with friends who randomly spend their time thinking about how you could be more successful, and do the same for them.
143
314
37
2,991
182,476
Adam Pollack retweeted
yoooo this is super interesting and aligns with some prior intuitions i had wrt reasoning length
New Anthropic Research: “Inverse Scaling in Test-Time Compute” We found cases where longer reasoning leads to lower accuracy. Our findings suggest that naïve scaling of test-time compute may inadvertently reinforce problematic reasoning patterns. 🧵
1
1
5
252
Our Researcher in Residence @yaboilyrical will be discussing his work on SMC steering at UC Berkeley on Aug 3. Check out the blog on this work here: nousresearch.com/steering-th… Details below!
excited to announce i’ll be a speaker at the @BerkeleyRDI DeAI Summit, presenting my research with @NousResearch on constrained language decoding! if you’re attending the summit, or want to come see my talk, please contact me!
9
7
114
17,309
Adam Pollack retweeted
excited to announce i’ll be a speaker at the @BerkeleyRDI DeAI Summit, presenting my research with @NousResearch on constrained language decoding! if you’re attending the summit, or want to come see my talk, please contact me!
1
4
2
36
19,038
Adam Pollack retweeted
i’m out of excuses to not be who i am
1
3
11
334