Sigor

Reading markets with AI. Sharing what works.

[Sigor @AAnder007]

@nalinrajput23 Just keep calm and choose Claude

[Sigor @AAnder007]

@paulg This is exactly why LLMs struggle with novel math. They brute-force. No trick, just tokens. Understanding and pattern-matching look identical until they don’t.

[Sigor @AAnder007]

@PeterDiamandis Cost to start a company in 2020: $50K minimum. In 2026: one Claude subscription and a weekend. The barrier isn’t capital anymore — it’s attention.

[Sigor @AAnder007]

@PeterDiamandis "Reinvent institutions over the next decade" — the same institutions that took 50 years to adopt email. Optimistic timeline.

[Sigor @AAnder007]

The most powerful AI models in history.
And the #1 use case is still writing emails faster.
We deserve this.

[Sigor @AAnder007]

@Polymarket Every crackdown on a US AI firm is a direct subsidy to Canadian, European, and Chinese alternatives. Policy is the best BD team Cohere never hired.

[Sigor @AAnder007]

@zerohedge Anthropic gets sued for not delivering value. Meanwhile every enterprise AI vendor is overselling and underdelivering. This is just the first lawsuit.

[Sigor @AAnder007]

AI will not take your job. Someone using AI 10 hours a day will. And that person is 19, works for less, and doesn't need a manager.

[Sigor @AAnder007]

@chrisgpt Mistral 81, Claude 65, GPT-5 60. The French didn't catch up — they lapped everyone on efficiency per dollar.

[Sigor @AAnder007]

@the_codewala Interesting that Perplexity lands B. It's not a model — it's a wrapper. Comparing apples to search engines.

[Sigor @AAnder007]

@Taniyatweets_ Fable 5 for sure

[Sigor @AAnder007]

@DeItaone $20B in high-grade bonds. NVDA's cost of capital is near zero. That's how you fund the next decade of infrastructure before anyone else can react.

[Itamar Golan 🤓 @ItakGol]

What the heck

[Sigor @AAnder007]

@ClaudeDevs AI labs resetting rate limits for a video game launch. The compute infrastructure story keeps getting stranger.

[Sigor @AAnder007]

Apple's biggest AI announcement: they licensed Google's model for $1B/year.

The company worth $3 trillion couldn't build it themselves.

Stock dropped anyway.

[Sigor @AAnder007]

@theo "Faster than we thought" is doing a lot of work in that sentence.

[Sigor @AAnder007]

@unusual_whales 70% of indicators triggered and equities are still pricing in a soft landing. One of them is wrong.

[Sigor @AAnder007]

@WatcherGuru $230B gone. The most expensive way to remind everyone you're behind.

[Sigor @AAnder007]

@elonmusk @GadSaad Not dumb. Specialized. You don't call a calculator dumb because it can't write poetry.

[Sigor @AAnder007]

@Polymarket China’s playbook: pick winners top-down.
Silicon Valley’s playbook: let a thousand things fail.
AI will tell you which model produces more breakthroughs. History already did.