I was also going to make a snarky comment about AI written blog posts but they are just so prevalent now it’s barely worth saying. I just have to suppress the irritation and parse out the information. And I’m guilty of doing AI documentation so I’m a total hypocrite.

He was the house curmudgeon - hilarious

Is this issue specific to 122B or will it lurk in other 3.5 r 3.6 models? I run a 9B model locally so while this fix doesn't help me I'm wondering if it's general?

As a bonus it will also repel horse flies.

https://www.science.org/content/article/zebra-stripes-confus...

Indeed, a lot of HN readers don't realise their good luck in life. It takes a dose of poverty to bring perspective.

I was taught 20% of customers generate 80% of the work. It seemed borne out by experience and I mostly fucked them off, the surprise/outrage when I told them they weren’t a good fit and told to go elsewhere was something to behold.

Slightly tangential to the article, which seems interesting, but the main issue with IPFS was the horrendous performance of clients which I seem to recall related to having a refresh storms, sparse routing tables, unreachable peers as well as lookup speeds. Mostly the reputation was so bad that people didn't bother with it, I dismissed it for my own project. If your only users are crypto-grift projects you're in a bad place.

True. But for how long? 6-9 months till Mythos equivalence in glm-5.4?

Depends if they can distill Mythos. They are creative, I think they'll hack a US company to get access to Mythos traces and do it.

The counter example is the administration’s great success in the Iran war. Oh, wait.

Shutup peasant, you'll get it when we say. And be grateful.