2024-09-16T14:23:10+00:00 | 🔗
@mhdempsey I think if you were good at prompting 4o, you actually still get more usefulness out of 4o. Much faster. I think it actually increases the gap though in the long run, because it's much harder to get a feel for the llm when you are interacting with o1 and long latency
2024-09-16T12:43:09+00:00 | 🔗
@LinLingBi Bruh
2024-09-16T02:39:07+00:00 | 🔗
i'm positive about o1 as saying something about the direction of AI and that there is somewhere to scale beyond data/compute
2024-09-16T02:10:07+00:00 | 🔗
Aside from the current R Ticket, every single living current and former President, VP and major party nominee is voting for Harris and Walz. Bushs, Cheneys, Pence, Romney, etc. ... https://t.co/HntVqtfUxq
2024-09-16T02:08:26+00:00 | 🔗
I think o1 models are really just (somewhat) closing the prompting gap. They are much more resistant to prompting and are set in their ways of thinking so if you don't know how to prompt it helps a ton but if you do....
2024-09-16T01:56:10+00:00 | 🔗
RT @harris_wins: BREAKING: In a shocking rebuke of Donald Trump and JD Vance, Republican Governor of Ohio, Mike DeWine, just eviscerated th…
2024-09-16T01:22:24+00:00 | 🔗
@wgussml @terrence_tao The hard part about this is that o1 is so damn slow, i would rather use 4o at some point
2024-09-16T01:09:15+00:00 | 🔗
Median voter found? https://t.co/oVLb9pcuXD
2024-09-16T00:34:33+00:00 | 🔗
RT @NateSilver538: I've been around these premises long enough that you hopefully won't mistake me for woke, but the Haitian immigrant stuf…
2024-09-15T23:13:24+00:00 | 🔗
@rauchg What about Infrastructure Weeks?