spectralquant's profile picture. quantitative researcher... math and coffee are my life

E. Z.

@spectralquant

quantitative researcher... math and coffee are my life

Literally can't do anything with claude anymore.

This is real Claude Code hits usage limits for the week in hours. I've never once hit their 5hr limit on the 20x plan I think this is the most egregious usage limit change thus far in the industry



E. Z. reposted

"You're telling me I'm not allowed to preach?" "I won't stop until Jesus tells me I can't preach God's word." While police harass a Christian preacher, a citizen reports a nearby crime. Cops say, "Call it in, we're busy!" With what? Silencing Christian faith?!


He who controls the Tylenol, controls the Universe


E. Z. reposted

🚨BREAKING: This beautiful 3 minute and 48 second compilation of Charlie's words of wisdom has gone viral. Rest in peace.


E. Z. reposted

Charlie was the best of us.


I'm going to scream tonight. Codex with High Reasoning *sucks* at coding. Everything has a try/except:pass block and you can never know what's failing. UGH. @openai you should fix the rewards on your code since it's really bad.


E. Z. reposted

I think @OpenAI just nerfed GPT5-Pro Here's a snippet from my latest message: "Thanks for sharing the project files — it looks like TokenCalculationService.swift is incomplete (it truncates at self.form...)" Even though my prompt was less than 60k tokens, it got ragged...


GPT5 Pro definitely turned a lot more stupid than it was over the past 2 days. What's going on @OpenAI? Why do you always do this switcharoo?


E. Z. reposted

It's Monday, survey the damage, cry a little, and start cleaning up

It's Friday, delegate your work to coding agent and start your weekend early



E. Z. reposted

introducing flowith NEO. the SOTA agent i built for the future. infinite steps. infinite memory. endless output. cloud-running for a year straight. benchmark-beating intelligence. zero cost. zero limits. enjoy.


E. Z. reposted

Realizing Trump told us to buy stocks and then paused tariffs setting up everyone vs. China

From Bags

E. Z. reposted

Retards would have you believe that we can build this but can't assemble an IPhone for less than 30,000 dollars.

bumbadum14's tweet image. Retards would have you believe that we can build this but can't assemble an IPhone for less than 30,000 dollars.

E. Z. reposted

BREAKING: Meta AI internally caught cooking benchmarks to hide poor results Absolute state.

ns123abc's tweet image. BREAKING: Meta AI internally caught cooking benchmarks to hide poor results

Absolute state.

llama 4 turns out to be a benchmark fraud lmao



E. Z. reposted

🚨NEW: Tim Dillon breaks down the current state of America and the offshoring of American manufacturing through the metaphor of Oompa Loompas, Willy Wonka, and Charlie and the Chocolate Factory. Absolutely BRILLIANT and HILARIOUS. (10 Minutes. MUST WATCH) 📸: @TimJDillon


E. Z. reposted

None of this "you are going to make everyone poorer and drop their standard of living" cope works because since I have been alive I HAVE WATCHED EVERYTHING I INTERACT WITH GET WORSE AND MORE EXPENSIVE. Google is unusable, its all advertisements. Every app is slow, and its UI…


E. Z. reposted

Oldies but goldies: S Shaobing Chen, D L. Donoho, and M A. Saunders, Atomic Decomposition by Basis Pursuit, 1998, Introduces l1 regularization for signal decomposition. epubs.siam.org/doi/abs/10.113…


The difference between @AnthropicAI's Claude and @OpenAI's ChatGPT? ChatGPT is the Redditor that argues with you and nitpicks at everything you say even if they're wrong. Claude is your bro trying to help you, he likes your ideas even if you expressed it imprecisely at first.


E. Z. reposted

Happy to announce we outperformed @OpenAI o1 with a 7B model :) We released two self-improvement methods for verifiable domains in our preliminary paper -->

yoshiyama_akira's tweet image. Happy to announce we outperformed @OpenAI o1 with a 7B model :)

We released two self-improvement methods for verifiable domains in our preliminary paper -->
yoshiyama_akira's tweet image. Happy to announce we outperformed @OpenAI o1 with a 7B model :)

We released two self-improvement methods for verifiable domains in our preliminary paper -->

E. Z. reposted

BREAKING: US Commerce Secretary Lutnick says Trump's goal is to abolish the Internal Revenue Service


Loading...

Something went wrong.


Something went wrong.