2024-09-17T13:56:17+00:00 | 🔗
Yuval Noah Harari talks about this on The Gray Area with Sean Illing... AI is the evolution of bureaucracy; AI is the evolution of the deep state. Safety is monitoring and control of AI, but also continuing to ensure that the people in control play by democratic rules.
2024-09-17T05:02:12+00:00 | 🔗
@jakeottiger No :(
2024-09-17T04:34:21+00:00 | 🔗
104 seconds now, asked it to make a lexer for me... https://t.co/XeKxcn1GEF
2024-09-17T00:24:35+00:00 | 🔗
Just today at work I found a prompt that is well within the scope of any LLM but is laughably impossible for any of the current LLMs to solve.
2024-09-16T22:56:44+00:00 | 🔗
@LinLingBi AHHHHHH
2024-09-16T18:52:47+00:00 | 🔗
@tszzl yes
2024-09-16T18:45:08+00:00 | 🔗
So we are in the "creating stories" stage of the election?
2024-09-16T15:54:47+00:00 | 🔗
My pet theory is that the reason why JD/Trump thought it would be believable that people are eating pets is that RFK Jr assured them that he knew lots of people (where some of them happened to be immigrants and in Ohio too) who ate pets.
2024-09-16T15:39:46+00:00 | 🔗
If you treated COVID like a normal sequence of DNA, and normal translation discretizing three base pair blocks to amino acids, you would have horribly fundamentally misunderstood the virus. Mistakes in AI + Bio come from thinking things are discrete or easily discretized. https://t.co/ueNHbQgB9D
2024-09-16T14:24:17+00:00 | 🔗
Anyone remember when time to first token was a big thing?