- nmcaIn a maximally earnest way — to what degree should we be sure these language harms are real on net? Are there data as opposed to anecdata? (Of course many phenomena are real without data; we can just be more confident in cases with data)
- This is also a nice way to combine the ratings of a number of noisy annotators with variable annotations noise.
- This is not work by any of the high profile new hires, in case folks are confused.
- Lillian’s blog is extremely good in general & if it’s new to you I suggest checking out the other posts also. I particularly enjoyed the one on human data.
- right, but you definitely shouldn’t be using any other formatter than ruff and this helps with that
- It’s the same principle as:
https://www.tripadvisor.com/Restaurant_Review-g60763-d477541...
- Isn’t the answer to the question just classic economies of scale?
You can’t run GPT4 for yourself because the fixed costs are high. But the variable costs are low, so OAI can serve a shit ton.
Or equivalently the smallest available unit of “serving a gpt4” is more gpt4 than one person needs.
I think all the inference optimisation answers are plain wrong for the actual question asked?
- You can dream of better yet! If the spec was required to be open source for the government project, then you could have commercial choices and some less feature rich open source version.
- I feel the AI safety community has not made enough of Lehrer’s masterpiece on the topic:
- Indeed, it’s like saying a jet plane can fly!
- It takes Tim Gowers more than hour and a half to go through q4! (Sure, he could go faster without video. But Tim Gowers! An hour and a half!!)
- It’s interesting that this is a competition elite enough that several posters on a programming website don’t seem to understand what it is.
My very rough napkin math suggests that against the US reference class, imo gold is literally a one in a million talent (very roughly 20 people who make camp could get gold out of very roughly twenty million relevant high schoolers).
- There wasn’t in the CoT for these problems.
- Great follow-up work from OpenAI on this:
- Why is the method called sendOnce? It’s send with a capacity limiter / semaphore right, so what about it is Once?
- This is great! I have daydreamed about how to do a more complex algo: instead of character speed and frequency you could do bi-character speed/frequency/error rate and probably improve over keybr further.
- “bad” seems fine?
- This article is quite bad: haughty in tone and confused in content.
If you are interested in learning something about a key part subject matter (compounding loopholes and their impact on drug prices), this article is much better:
- How do you purport to know this?