She has 7.5k+ citations and expertise in NLP/LLMs, yet she still had to thoroughly revise everything and go through all the coding-test shenanigans. It really is a dog-eat-dog world.
I quoted a boomer $40 to mow his lawn, edge it, and bag the clippings.
He laughed and offered $12 "like the neighbor kid used to charge." The neighbor kid charged that in 1994. Gas alone for the mower costs more than that now. He hired someone else. They quit after one mow. He's still looking.
Boomers need to go.
@burkov@Zai_org https://t.co/2xT7uDgOlX has both kimi and glm 5.2 around the 0.5/0.05/2.25 area. No shot you're telling me that you will get more tokens out of a sub for the same price 😭 (I use both)
Imagine being forced to give someone 6.2% of your paycheck, every single check, every month, for 30 to 50 years of your life.
And that person said, "Don't worry, I'm holding this for you and will pay it back to you on a monthly basis, when you retire at 65."
And then they said, "Nah, just kidding. I meant when you retire at 67. And at that time, I'll only give you 70% of what you paid me."
And then they said, "Oops, I spent all your money. You're out of luck."
That's the U.S. government.
Stanford's public course lectures helped develop my understanding of ML and NLP technologies, where my university hadn't been able to catch up to quite yet.
“I started my process by watching all the lectures from Stanford’s Language Modeling from Scratch course, which is helpful for illustrating the breadth of topics I needed to learn and helped me organize many scattered concepts in my brain into one coherent picture of the field.”
I’m starting to feel like GLM 5.2 might be better than Opus.
I was spending $300/day on Claude. Switched to GLM, spent $3.82 today, and it found and fixed a Claude bug from yesterday.
I honestly can’t tell which is “better” anymore.
I pay for a codex/chatgpt plus sub and use https://t.co/sEqF6PJvTl for anything programmatic or for less important tasks since they have the lowest cost glm 5.2 and kimi k2.7 api prices. I also have a claude pro sub, but I don't count it since it was paid for by my university for research purposes.
Bernie Sanders introduces a bill to give US citizens a 50% ownership stake in ‘large’ AI companies
It is estimated every citizen will be paid more than $1,000 annually
Say I’m interested in investing in open source startups what is a realistic amount of money to start with?
With stocks you can kind of start small but then again I don’t care about stocks and established companies.
I’m not great at financial decision making lol
I wish I could be excited but i know this model is not going to drop cheaper than flash 3.5, and that model is bad at everything.
Google has to have some sort of miracle to make 3.5 pro preferable and cost effective..
Gemini 3.5 Pro leak 👀: it will be released this week
Google said Gemini 3.5 Pro would arrive in June after delaying it at I/O, and all signs are pointing to a launch this week.
What to expect:
> Major coding improvements
> Better agent workflows
> Stronger UI and frontend generation
> 2M+ context window
> More aggressive pricing than GPT-5.5 and Fable 5
The timing is interesting
Fable 5 is fighting regulators amd GPT-5.6 is still nowhere to be seen
And Google might be about to drop its most important model of the year
This week could get very interesting 👀
do you think Gemini 3.5 pro will beat Fable 5?
Gemini 3.5 Flash is worse than GLM 5.2 while costing roughly 2x as much
It was released a month ago. Google needs to wake up because right now they're dead, they don't lead a single benchmark anymore (except video gen, but not for long)