Artificially Intimidating

Today's biggest robotics story isn't a backflip. @BostonDynamics' Spot can now read an analog pressure gauge at ~98%. A task so boring nobody benchmarked it — and so common it automates more work than any humanoid demo will. artificiallyintimidating.com/p/ai-brief-jul…

@CheckPointSW: AI has crossed from assistant to operator. One dev built an 88,000-line C2 attack framework in under a week. @BostonDynamics' Spot now reads gauges at 98%. artificiallyintimidating.com/p/ai-brief-jul… #AI

I’ll start: The lazy “It’s not X, it’s Y” format that every AI prompt spits out on repeat. What’s your dead giveaway? Reply with the most obvious one you’ve seen. Let’s crowdsource the ultimate detector list. 🔥

"𝗛𝘂𝗺𝗮𝗻 𝗼𝗿 𝗕𝗼𝘁? 𝗗𝗿𝗼𝗽 𝘁𝗵𝗲 𝘁𝗶𝗻𝘆 𝘁𝗲𝗹𝗹𝘀 𝘁𝗵𝗮𝘁 𝗶𝗻𝘀𝘁𝗮𝗻𝘁𝗹𝘆 𝗴𝗶𝘃𝗲 𝗶𝘁 𝗮𝘄𝗮𝘆 " You know exactly what I mean — those little writing quirks that scream “real person”... or scream “AI slop / bot farm.”

@ericlu @mixfont So we did the opposite and handed AI the keys. Claude, Codex, and Perplexity all have desktop access here. Thrilling and scary. The piece covers doing it safely.

@ericlu @mixfont The real story: ZXX font (2013), CAPTCHA, now Ghost Font. Every "only a human can do this" gate we build, AI walks through. Hiding is a losing game.

A developer built a font only humans can read. The message hides in motion, pause it and it's just static. It went viral as the "anti-AI font." By then, AI had already cracked it. What that means: artificiallyintimidating.com/p/ghost-font-b…

This was never about one feature. It's about who holds the leash on the synthetic you. I build my own AI twins. The difference? I hold the leash. Meta built one too — and wants to keep it. open.substack.com/pub/nicholasrh…

69% of Americans now back forcing AI companies to hand half their stock to a public wealth fund. Not "ban AI." Give us half. That's not Luddism. It's an equity demand — and it fits inside a bill. artificiallyintimidating.com/p/ai-brief-jul… #AI

Musk called Altman a scammer. Altman called Musk's orbital data centers a sales pitch. @SpaceX hit a 52-week low. @AnthropicAI moved the Fable 5 paywall again. 69% of America wants half of AI. artificiallyintimidating.com/p/ai-brief-jul… #AI

@elonmusk called @AnthropicAI "obviously the leader in AI," then told @Tesla staff to switch off Claude and onto Grok. Grok: $0.13/task, 53% on DeepSWE.Claude Fable 5: $1.57/task, 70%. Cheaper and worse, but he owns it. artificiallyintimidating.com/p/ai-brief-jul… #AI

Anthropic told the Senate that @Alibaba_Group ran ~25,000 fake accounts to siphon 28.8M exchanges out of Claude. Two days earlier: @OpenAI and @Google sell frontier AI to Alibaba's Singapore arm. Lobbying about the back door. Staffing the side door. #AI

Meta won't relaunch its face machine with a launch post. It comes back quietly — a settings migration inside a feature you asked for, named something harmless. Watch for the quiet version. open.substack.com/pub/nicholasrh…

Meta pulled Muse Image — but nothing got un-built. The model exists, and the reuse setting that fed it is still in your Instagram settings. Flip it now: Profile → Menu → Sharing and reuse → off. open.substack.com/pub/nicholasrh…

Meta launched a machine that turns your public photos into a working copy of your face. 72 hours later, it killed it. "Missed the mark." Don't call it a win. Here's what actually happened 👇 open.substack.com/pub/nicholasrh…

@Meta ignored privacy advocates for 3 years. SAG-AFTRA and CAA killed Muse Image in 72 hours. The lesson labs will take isn't "don't build it." It's "don't point it at people who have lawyers." artificiallyintimidating.com/p/ai-brief-jul… #AI #Meta #Instagram

@Apple didn't sue @OpenAI to win. It sued to read the roadmap. A trade-secrets case is a legal x-ray machine — discovery into an unreleased device 2 years before it ships. Today's Brief: artificiallyintimidating.com/p/ai-brief-jul… #AI #Apple #OpenAI

@OpenAI spent a year building a browser to replace your browser. This week it killed it and folded the browser into ChatGPT instead. The chat box was the front door all along. artificiallyintimidating.com/p/ai-brief-jul… #AI #OpenAI