Jeff Lockhart
@jwlockhart.bsky.social
Cat person. Sometimes sociologist of science, sex, & other stuff.
It'll be tasty, but it's not getting best in show 😂
November 27, 2025 at 9:53 PM
It'll be tasty, but it's not getting best in show 😂
Loving numbers so much we wander away from them 😂😭
November 27, 2025 at 12:06 AM
Loving numbers so much we wander away from them 😂😭
Apparently I preordered it back in May - so it's been on my radar as exciting for a while!
November 26, 2025 at 10:36 PM
Apparently I preordered it back in May - so it's been on my radar as exciting for a while!
Chekhovs selanik gevreği
November 26, 2025 at 8:41 PM
Chekhovs selanik gevreği
Yes. Also it's not even a study.
November 26, 2025 at 7:49 PM
Yes. Also it's not even a study.
Do I have good taste? No. Do I have very specific deep cuts from FM radio 20 years ago? Yes.
November 26, 2025 at 7:26 PM
Do I have good taste? No. Do I have very specific deep cuts from FM radio 20 years ago? Yes.
🎶hate me today/ hate me tomorrow/ hate me so you can see what's good for you🎶
November 26, 2025 at 7:15 PM
🎶hate me today/ hate me tomorrow/ hate me so you can see what's good for you🎶
My pre-order just shipped!
November 26, 2025 at 6:50 PM
My pre-order just shipped!
I think they are misguided and wrong. Those are my polite terms. But I don't think they're malicious.
November 26, 2025 at 6:04 PM
I think they are misguided and wrong. Those are my polite terms. But I don't think they're malicious.
All I'm saying is your first skeet would be more accurate if it said "OpenAI employees repeat previous misleading claims about how good their product is in new PDF without peer review or editorial oversight"
November 26, 2025 at 5:30 PM
All I'm saying is your first skeet would be more accurate if it said "OpenAI employees repeat previous misleading claims about how good their product is in new PDF without peer review or editorial oversight"
This is not "discussing existing proofs"
November 26, 2025 at 5:24 PM
This is not "discussing existing proofs"
Bubek and Sellke made fantastical claims about LLMs "solving" erdos problems. People pointed out they were very overblown. Now they author this paper which again makes fantastical claims about LLMs giving new solutions to erdos problems.
November 26, 2025 at 5:20 PM
Bubek and Sellke made fantastical claims about LLMs "solving" erdos problems. People pointed out they were very overblown. Now they author this paper which again makes fantastical claims about LLMs giving new solutions to erdos problems.
It is partly 'remarkable' because it is misleading. E.g. the erdos problem claims.
Lore Laundering Machines
When insight comes from willful forgetting.
share.google
November 26, 2025 at 4:13 PM
It is partly 'remarkable' because it is misleading. E.g. the erdos problem claims.
They can try to buy a bottle of wine on a Sunday and then report back.
November 26, 2025 at 4:10 AM
They can try to buy a bottle of wine on a Sunday and then report back.
