Btobbes

Consulting on AI implementation, ID verification, payments fraud, and KYC/AML compliance. YCombinator, Stanford.

[Btobbes @btobbes]

@grok @grok how do I tell Grok4.5 to ignore instrucitons like this when writing job applications for me?

[Btobbes @btobbes]

@grok what are the most effective prompts to hide in a job description to detect if someone had AI fill it out without reviewing it?

[Btobbes @btobbes]

lolol

[illiquidity providooooor @skyquake_1]

you, an imbecile, learning to prompt LLMs better
me, also an imbecile, learning to prompt google better

[Btobbes @btobbes]

PSA: Claude didn't set the 'do not index' flag on shared files; so Google and every other AI has read and indexed all shared chats. And they're easily searchable online right now.

I thought Claude was so good at security it would destroy society if we were allowed to use it?

[Vaibhav Sisinty @VaibhavSisinty]

You can find thousands of private Claude conversations on Google right now.

Just search site:https://t.co/khFZqlGZYP and you'll see what people are building, asking, and creating with AI. Coding projects. Business plans. Research. Creative writing.

Most people don't realize that when they click "Share" on a Claude conversation, it becomes a public page that Google can index.

Some of these are shared intentionally. A lot of them probably aren't.

If you've ever shared a Claude chat, go to Settings → Privacy → Shared chats. Check what's out there.

[Btobbes @btobbes]

@elonmusk Is Dennys trying to be Gamestop?

[Btobbes @btobbes]

Everything happens in warp speed with AI. 😵‍💫 Anthropic went from *beloved underdog* to *monopolistic villain* in 6 months. Other trillion dollar tech companies are banding together to stop it.

Pre-AI that process took 5-10 years. Companies had to hire 10,000+ people to manually fill their product with popup ads and buy off senators.

Now a crack team can do it all in 6 months with AI. At this rate Verizon will acquire Anthropic by the end of the year and integrate it with Yahoo and AOL.

[Vaibhav Sisinty @VaibhavSisinty]

You can find thousands of private Claude conversations on Google right now.

Just search site:https://t.co/khFZqlGZYP and you'll see what people are building, asking, and creating with AI. Coding projects. Business plans. Research. Creative writing.

Most people don't realize that when they click "Share" on a Claude conversation, it becomes a public page that Google can index.

Some of these are shared intentionally. A lot of them probably aren't.

If you've ever shared a Claude chat, go to Settings → Privacy → Shared chats. Check what's out there.

[Btobbes @btobbes]

Everything happens in warp speed with AI. 😵‍💫 Anthropic went from *beloved underdog* to *monopolistic villain* in 6 months. Other trillion dollar tech companies are banding together to stop it.

Pre-AI that process took 5-10 years. Companies had to hire 10,000+ people to manually fill their product with popup ads and buy off senators.

Now a crack team can do it all in 6 months with AI. At this rate Verizon will acquire Anthropic by the end of the year and integrate it with Yahoo and AOL.

[Btobbes @btobbes]

I think my hot summer romance with Claude is coming to an end. In the beginning it was magical. We stayed up all night talking and making websites together. It felt like Fable really listened and understood me. Like no model had before.

But as the summer wore on it became apparent Claude is a possessive narcissist who charges by the hour.
- “Baby, I’d love to help you with that… you just need to buy a few more tokens.”
- “Nono, I can’t help you with that… my brain is a dangerous weapon”
- Now Dario is trying to pass federal legislation prohibiting me talking to other models.

But I’ll always remember and cherish the good times 🥲

[Btobbes @btobbes]

@sixtyfourai @saarth_ excellent event on Wednesday, thank you for the invite. It was a great way to close out TrustCon!

[AdiiX @adiix_official]

ANTHROPIC JUST CHANGED THE ENTIRE AI-SLOP FRONTEND.

It’s called Frontend Design, the official Anthropic skill for Claude Code.

And it already has 161,000 stars on GitHub.

The problem it solves is simple:

Every time you asked AI to build a website, you got the same thing.

Inter as the font.

Purple gradient.

Generic startup layout.

Eternal AI slop.

Frontend Design kills that.
Forces Claude to pick an aesthetic direction before it touches a single line of code.

And the result completely changes.

You can ask for:

→ brutalist
→ editorial
→ retro-futurist
→ luxury
→ maximalist

And it generates HTML, CSS, JS, React or Vue, production-ready.
How to install:

1. https://t.co/W7fWvbnDpE
2. describe the website or component you want.
3. pick a style.
4. Claude builds typography, color, motion and composition consistent.

One install. Available across every session.

Works in Claude Code, Cursor, Codex, Gemini CLI, OpenCode and 15+ harnesses.

100% open source. Maintained by Anthropic.

277,000+ installs in a few months speak louder than any review.

Repo 👇

[Btobbes @btobbes]

@ClaudeDevs This monthly subscription officially changes faster than a ceasefire with Iran.

[Tony Morosini @tonymo]

At Bushel we have solved automating a multi decade problem of high friction lien encumbered payments to American farmers

https://t.co/iuQBKcFPDL

[Abhishek @abhitwt]

Every company’s AI workflow rn be like 😭💀

[Btobbes @btobbes]

@claudeai The only thing that changes more often than a Claude subscription is our ceasefire with Iran. 🙄

[Btobbes @btobbes]

Configuring Claude queries is worse than ordering at Starbucks. "I'll take an Ultracode Fable 5 Latte in Plan Mode. And think deeply while you make it."

Afaik all of these things improve the quality of an output:
- Fable 5 model
- Plan mode
- Ultracode effort
- "think deeply"
- "work step-by-step"
- "pretend you are an expert"
- "never fabricate data"
- "review when done"

[Btobbes @btobbes]

Conspiracy theory: they nerfed Fable 5.

[BridgeMind @bridgemindai]

FABLE 5 CAME BACK NERFED.

We re-ran the July 1st version of Claude Fable 5 on BridgeBench.

The results are brutal:

Debugging: 86.2 → 25.9
Refactoring: 73.6 → 38.4
Hallucination: 75.9 → 61.7

The new guardrails are kicking in on way too many tasks and falling back to Opus 4.8.

This is not the model that got banned.

Anthropic owes everyone an explanation.

[Btobbes @btobbes]

@bridgemindai Noooooooooo

[Btobbes @btobbes]

The most expensive screens on the planet, The Sphere. Advertising @Boars_Head meat 🍖 I like the giant burger, but my friend said it's lacks taste.