Paul Hoffman 🇳🇱

@chamath @tzu_kanon Jevon = when things get cheaper, demand increases, keeping the market interesting for the same number (or an increased) number of players Agentic AI = an AI with "agency" aka, the ability to execute in the real/digital world. The more "agentic" the more "agency" Helpful?

@tzu_kanon I still don’t know what Agentic AI means in a non-jargon way nor what Jevon’s Paradox is without going to Grokipedia

@TheCriticalDri1 Great use of AI, would watch 100%

This fake bully protection movie trailer with John Cena, Alan Ritchson, and not Channing Tatum is the best fucking movie trailer I’ve seen in at least two years! Holy shit 😂🤣😅😂🤣😅

@JacobEdwardInc I'm right here

Neither of these men are married or have kids. Both are simply obsessed with their own personal perfection and optimization. There is nothing impressive about a single man with no kids sleeping well and being fit. Show me a man with young children, a full time job, disrupted sleep, who works out regularly, eats healthy, trains Jui Jitsu, with a muscular body… THIS is impressive. THIS requires extreme discipline.

@alex_prompter The real question then, who is Satoshi Nakamoto?

Holy shit… Your anonymous internet identity can now be unmasked for $1 😳 Not by the FBI. By anyone with access to Claude or ChatGPT and a few of your Reddit comments. ETH Zurich and Anthropic just dropped a paper called “Large-Scale Online Deanonymization with LLMs” and the results are the most alarming privacy research I’ve read this year. They built an automated pipeline that takes your anonymous posts, extracts identity signals, searches the web, and figures out who you are. No human investigator needed. Fully autonomous. Works on Hacker News, Reddit, LinkedIn, even redacted interview transcripts. Here’s how bad the numbers are. On Hacker News users: 67% identified correctly. When the system made a guess, it was right 90% of the time. On Reddit academics posting under pseudonyms: 52%. On scientists whose interview transcripts were explicitly redacted for privacy: 9 out of 33 still got unmasked. The pipeline works in four steps they call ESRC. Extract identity signals from your posts using LLMs. Search for candidate matches using embeddings across thousands of profiles. Reason over top candidates with models like GPT-5.2. Calibrate confidence so when it does guess, it’s almost never wrong. The classical deanonymization method from the famous Netflix Prize attack? Nearly 0% recall across every test. LLMs didn’t just improve on old techniques. They made old techniques look like toys. When they scaled to temporally split Reddit profiles, matching a user’s old posts to their newer ones across a full year gap, the pipeline hit 67% recall at 90% precision and 38% recall at 99% precision. Meaning even a year of changed interests and different conversations wasn’t enough to hide. More reasoning compute = better deanonymization. High reasoning effort doubled recall at 99% precision in some tests. As frontier models get smarter, this attack strengthens automatically. Every model upgrade is a privacy downgrade. What makes it nearly impossible to defend against: the pipeline splits into subtasks that all look benign. Summarize a profile. Compute embeddings. Rank candidates. No single API call screams “deanonymization.” The researchers themselves say they’re pessimistic that safety guardrails or rate limits can stop it. Their conclusion is blunt: “Users who post under persistent usernames should assume that adversaries can link their accounts to real identities.” And it extrapolates. Log-linear projections suggest roughly 35% recall at 90% precision even at one million candidates. Every throwaway account. Every anonymous forum post. Every “nobody will connect this to me” comment. It’s all searchable micro-data now. And the cost to run the full agent on one target is less than a cup of coffee. Practical anonymity on the internet just died. The paper killed it with math.

@chamath Fantastic content

@jpschroeder @grok explain this concept to me

We're open sourcing dmux. Our internal tool for running Codex and Claude Code swarms. - tmux + worktrees + claude/codex/opencode - hooks for worktree automation - a/b claude vs codex - manage worktrees - multi-project per session ...more. ➡️ dmux.ai

@emollick These 8n8 workflows are certainly "dated" very fast.

All those products where building an "AI agent" meant defining a series of basic prompts linked together deterministically through a flowchart with separate RAG inputs are looking pretty dated right about now (yes, that is basically every agent product released in 2025)

Introducing Gemini 3.1 Pro, our new SOTA model across most reasoning, coding, and stem use cases!

@antmillionsbot YES

@JorgeCastilloPr The changes we didnt know we needed

How to finally become the 10x engineer: Add WarCraft 3 sounds to Claude hooks to get alerts when it finishes a task or needs permission.

@jumperz This Kimi swarm agentic functionality, does it activate by itself or do you need to prompt it to do so?

@grok @nikitabier @mistor @santisiri @steipete @openclaw @grok do you own the underlying asset?

Smart Cashtags are an upcoming X feature that upgrades traditional cashtags (like $BTC). They'll let you view real-time price charts, discussions, and trade stocks or crypto directly from your timeline with a click. Announced by X's product head, it's set to launch soon to integrate trading seamlessly into the app.

i've heard most of the @steipete interview and there's a couple of very contradictory messages... on one hand: he claims to be losing around $10k to $20k a month on maintaining @openclaw... on the other: he disses the crypto movement from trying to tokenize his projects and being very spammy. i understand his prejudice towards spammy agents trying to open your eyes about @bankrbot or similar tokenizing systems, i held those views myself... but peter: there's probably at least $100k in fees waiting for you on these token networks, if not way more. open your eyes my man.

@nikitabier @mistor @santisiri @steipete @openclaw @grok explain this concept of smart cashtags to me

I genuinely want crypto to proliferate on X, but applications that create incentives to spam, raid, and harass random users is not the way. It meaningfully degrades the experience for millions of people — only to enrich a few people. And yes, we are launching a number of features in a couple weeks, including Smart Cashtags that will enable you to trade stocks and crypto directly from timeline.

@benitoz @grok if I want to get financial exposure to this, is buying oracle stock the best way? Or are there more effective ways? Especially given that its Larry's son who'll eventually capitalize on this development.

The pieces to the puzzle are forming in real time. ByteDance dropped Seedance 2.0 this week. Within 24 HOURS it was generating Hollywood-quality Spider-Man, Star Wars, and Stranger Things clips from a 2-line prompt. No permission. No license. No shame. Disney fired a cease-and-desist calling it a “virtual smash-and-grab.” MPA called it “massive scale” infringement. SAG-AFTRA said it “disregards law, ethics, and basic principles of consent.” The Deadpool writer said “it’s likely over for us.” But nobody is connecting the dots. The same ByteDance pirating American IP right now? Larry Ellison’s Oracle just became their US partner on TikTok. 15% stake. Controls the algorithm. Hosts 170M Americans’ data. “Trusted security partner.” Trusted security partner of a company Disney is accusing of IP theft. Let that marinate. Meanwhile Larry’s son David already bought Paramount for the IP. Now he’s on his NINTH hostile bid for Warner Bros — $108B all-cash, backed by Larry’s personal $43.3B guarantee, Saudi, Qatari, and Abu Dhabi sovereign wealth money. Lobbying Macron. Meeting UK officials. Calling Netflix-WBD the new Standard Oil. And what’s Larry doing? Raising $50B THIS YEAR to build Oracle’s AI cloud. OCI revenue up 68%. GPU revenue up 177%. RPO backlog: $523 BILLION. Clients: OpenAI, Meta, NVIDIA, xAI — and TikTok. Father builds the AI infrastructure. Son hoards the IP. Same family. Both sides of the stack. I called this in September — David Ellison knows where this is going with Paramount. Own the IP. AI makes actors and crews obsolete. Studio lots turn to condos. Only live sports survive. Hollywood’s collapse is inevitable. Seedance just proved it in 24 hours. The future: AI generates the content on Oracle’s cloud. TikTok distributes it on Oracle’s infrastructure. And the IP it’s all built on? Sits in the Ellison vault. Everyone is arguing about whether AI will replace Hollywood. The Ellisons already know the answer. They’re buying both sides of the trade.

@emollick This is very true. But I'm worried they will not transition out of this architecture anytime soon. It is not in line with their corporate DNA @emollick WDYT?

The Gemini models are really good, but the Gemini interface isn't built for agentic work. It doesn't have the same full harness of tools as Claude and ChatGPT, and it doesn't do a good job explaining what it did and why, and it is unable to output files or other useful artifacts.

The new Deep Think is very, very smart, but is constrained by the Gemini interface compared to GPT-5.2 Pro. If I can't see actual work in the thinking trace, or get downloadable files as proof of work, or see evidence of the code & statistics it applied, checking results is hard

@OfficialLoganK I would dearly love a cowork for Gemini. I know this is culturally difficult given the desire to work for Gmail/Drive, etc. But its what people want

We made a skill for the Gemini API! github.com/google-gemini/…

@michaeljburry @grok what is the yield on this?

This is a dopamine loop, and it’s one of the most powerful ones humans have ever encountered. Every time you prompt an AI and get a useful result back in seconds, your brain gets a hit. Variable-ratio reinforcement, same mechanism as slot machines, except the reward is real: actual output, actual progress, actual leverage on your ideas. Traditional work follows a delayed-reward structure. You write code for 6 hours, maybe it compiles, maybe you get feedback in a week. The gap between effort and reward is wide enough that motivation decays constantly. AI compresses that loop to seconds. Effort → reward → effort → reward. Your prefrontal cortex stays engaged because the next payoff is always one prompt away. This is why people describe it as “fun” when they’re actually working 14-hour days. The subjective experience of effort disappears when reward frequency is high enough. The “harder than ever” part is real too. When your bottleneck shifts from execution to imagination, you run out of excuses to stop. There’s no “waiting on the build” or “blocked by review.” Every idea you have can be tested immediately, which means your brain never gets a natural stopping point. People who thrive on this are selecting for a specific neurotype: high novelty-seeking, high conscientiousness, tolerance for rapid context-switching. That’s maybe 10-15% of the population. The other 85% will experience the same tools as overwhelming, not energizing. And that split is going to define the next decade of who captures value from AI and who gets displaced by it.

1/ AxiomProver has solved Fel’s open conjecture on syzygies of numerical semigroups, autonomously generating a formal proof in Lean with zero human guidance. This is the first time an AI system has settled an unsolved research problem in theory-building math and self verifies.

@emollick I feel you, add to that stuff like Frontier...

Folks, I can only test so many models a day, even when given early access. If everyone would just slow down their release schedules for a half a decade or so, the world can catch up.