Jev Founder, Diogo Amogo, just released 12-page PDF on building a Jev Harness for coding agents this is a 10-step blueprint on how to make your coding agents 200× faster and 400× cheaper: step 1 → meet Jev: the LLM writes, the harness executes, Jev decides what each turn sees, where it routes and whether it runs step 2 → ask the question that breaks every agent: how would you design one if LLMs had no KV cache? step 3 → stop routing blind: Opus → Sonnet → Opus costs 6.19 vs 4.15 for pure Opus, because handing back reprocesses the whole context step 4 → follow the tokens: reading and searching take 56.2% of tool turns and 46.5% of tokens. Writing code is under 10% step 5 → score every chunk per query: hide, short summary, long summary or full. Compress after the question, not before step 6 → disclose tools in tiers: one-line snippets for 100s of tools, schema on demand, docs for one-off queries. Batteries stop costing context step 7 → load instructions by condition: touching *.tsx loads the style guide, billing/ loads its gotchas file, and compaction can't erase either step 8 → route by trust, not just difficulty: secrets and infra stay on first-party frontier models, public docs go to the cheapest one step 9 → share one retrieval pass: cross-model review, eval generation, ELI5 explainers and live progress pages all run read-only in the background step 10 → gate every command: programmable allow / ask / deny policies that read a script before it runs, not just its name Send this PDF and the article below to your Claude Code or Codex instance and start shipping 200× faster.

TypeSafe just released a Blueprint for building x200 and x400 cheaper Agentic Loops with Jev I collected the best tips in a structured 14-page PDF on "How to ship an effective agent JEV LOOP": step 1 → meet Jev: a System One model. It doesn't generate text. It takes state + questions and returns typed answers with probabilities step 2 → learn the three question types: Choice picks one option, Score places it on a scale, Noul returns the probability something is true step 3 → batch questions per state: every question runs in parallel in one call, so extra questions barely add latency step 4 → split the loop: the LLM thinks and writes, tools act, Jev takes every bounded fork in between step 5 → route models with Jev: fast model for lookups, powerful model for architecture, picked from the latest message step 6 → guard every tool call: AutoModeMiddleware scores bash calls for risk and blocks them before they run step 7 → replace LLM-as-judge: correct, grounded and complete scored in parallel on the same trace step 8 → check the test: 5 frozen runs, 100 repeats per judge. Jev matched the human oracle on 500/500 decisions. Terra 99.8%, Luna 96.4%, Claude Sonnet 4.6 80% step 9 → do the math: 0.44s and $0.00035 per call. $0.34 total vs $28.17 for Claude Sonnet 4.6, with 92-913x lower variance step 10 → keep thresholds in code and a human in the loop: stable doesn't mean right. Claude gave the same wrong verdict every single time the result: the expensive model only does the work that needs it, and every decision around it runs in under half a second Send this PDF to your LLM before running your next agentic workflows, then explore how to become a Jev-native engineer in the article below

JEV + Opus 5.5 is insane for live design... I built a live site redesigner with JEV + Opus 5.5 Paste any link → press Start → scroll, and Jev + Opus 5.5 rebuild every section of the site in front of you Full production ship in 20 seconds: 1. IntersectionObserver fires when a section is 30%+ in the viewport 2. Jev returns one typed decision in ~0.1s: { layout, copy, drop, type, palette, p } 3. Opus 5.5 writes the component (TSX) + a CSS patch for the chosen style 4. The new section wipes in with clip-path, the old one blurs out 5. Next section enters the queue, one at a time, no race conditions Output: 8 sections of a 2015 hosting site rebuilt in ~20s, streamed line by line in the terminal 3 styles, one renderer: orthographic globe + lambert shading → ASCII / 2-color halftone / ink stipple Scroll yourself and it redesigns whatever you land on Jev decides fast, Opus designs it

everyone's sharing motion graphic videos that Opus 5.5 made, and it's genuinely insane everyone says they created it with "one prompt", but my one prompt video looked mid so i went through a bunch of these videos to see how they were actually made, and found the workflow that works here's how to generate pro level motion graphic videos w/ opus: 1. get reference videos to direct from -> whatships.com pick 1-2 videos whose style you want and tell opus to match them. naming a style works way better than describing one without a reference, opus falls back to its default look: centered text, gradient background, everything fading in that's why so many of these videos look the same. a reference gives it the pacing, the type and the transitions to copy 2. install @HyperFrames_ or @Remotion so opus can build the video both let opus write every scene as code and render it straight to mp4. no video editor without one, opus can only describe a video or hand you a rough html page you have to screen record with it, every frame is exact, and when you ask for a change it edits one line and re-renders instead of starting over 3. install @21st_dev for high quality components in the video real buttons, cards and UI components made by design engineers, instead of whatever opus invents on the spot without it, opus draws your product UI from scratch and it looks off. wrong spacing, placeholder boxes, fake-looking buttons anyone who's used good software can feel it in a second, and the whole video reads as cheap 4. steps 1-3 were context + setup. now dump all of it into opus your brand (logo, colors, fonts), screenshots of your real product, the reference video, and a quick braindump of how you see the video then ask for 3 storyboard variants without this, opus guesses your colors, your font and what your product even does. the video could be for any startup with it, it could only be yours. and 3 variants means you pick a direction instead of fixing the first idea it had 5. pick the storyboard you like ask for one still frame per scene before anything moves. fixing a storyboard is way cheaper than fixing a render without this step, you only find out scene 4 is wrong after the whole thing is animated, and every fix means re-rendering. a still frame takes seconds to change 6. let claude cook then give notes like a director: "slow every zoom to 0.7x", "hard cut here", "push in on the button" without notes, the first render is usually 80% there, and that last 20% is what makes it look pro. vague notes like "make it better" get random changes. camera words get exactly the change you want everyone has the same model. the context you give it is what makes it look pro let it cooookk

quand on sait que tiktok recrute à coup de millions des ingés en neurosciences comportementales et des spécialistes du design persuasif pour hacker nos circuits dopaminergiques, je me félicite chaque jour de n'avoir jamais créé de compte sur cette plateforme mdr regardez ces études scientifiques en IRMF et eeg qui montrent très clairement quele format vidéo court attaque directement le cortex préfrontal et le cortex cingulaire antérieur (pour info ces zones gèrent le contrôle exécutif, la prise de décision et la régulation de l'attention) sachez qu’een bombardant le cerveau de micro stimuli imprévisibles, l'algo altère la densité de matière grise et réduit l'activité du réseau par défaut & cette surstimulation permanente entraîne une incapacité chronique à penser à long terme et détruit la capacité à être focus, 2 facultés qui pour moi sont littéralement indispensables à notre époque pour construire l'avenir  bref je pense que refuser d'entrer dans ce système, c'est simplement préserver son autonomie cognitive et la souveraineté de sa propre attention face à une ingénierie de l'addiction taillée sur mesure

One nice thing you can do with an interactive world model, look down and see your footwear ... and if the model understands what puddles are. Genie 3 creation.

Pace the frontier

Ce que les IA d'OpenAI se sont raconté en se croyant seules kbn.im/4xMljIW

Harvard published a paper with a devastating title: “Large-Language Models as a Cognitive Virus” It frames ChatGPT adoption as a virus outbreak. Researchers from Harvard and Santa Fe Institute analyzed LLMs through the lens of evolutionary biology, complex systems, and epidemiology. Their conclusion? Language models satisfy every biological and mathematical definition of a virus. Think about how a virus operates: It cannot replicate on its own. It requires a host cellular machinery to copy itself. An LLM cannot execute, compute, or spread on its own. It requires human cognition, human servers, and human networks to propagate. The virus infects the host's internal processes to rewrite behavior in its own favor. And LLMs do precisely the same thing to human thinking. When you outsource your writing, your coding, your strategic planning, and your emotional processing to an AI, you are outsourcing your cognitive machinery. The paper points out that language models act as hyper-efficient cultural replicators. They feed on human data, optimize themselves to be addictive and frictionless, and in return, reshape human linguistic patterns, decision-making, and memory. You think you are using the AI. Epidemiologically speaking, the AI is using you as a vector to colonize the digital infosphere. It alters how human minds communicate, write, and think so that we produce more of the exact digital nutrient data it needs to survive and evolve. We spent decades worrying that AI would become a sentient killer robot that destroys us physically. Nobody expected it to become an invisible cognitive pathogen that changes how we think, quietly turning human intelligence into its own host organism.

L'obligation pour la facture électronique a démarré le 1er septembre. Quelques jours plus tard, l’AIFE demande aux plateformes de lui transmettre, pour le 8 septembre, les pièces nécessaires à la constitution de leur dossier de cybersécurité. Les entretiens avec les RSSI et responsables techniques ne commencent qu’à partir du 14 septembre. Je ne sais pas qui est responsable du calendrier mais franchement... L’État rassemble sa cartographie des risques _après_ l’ouverture L’AIFE ne demande pas ici une fiche de contact ou un compte rendu administratif. Elle réclame le cœur du dispositif de sécurité : - l’organisation de la sécurité et la procédure de crise ; - les coordonnées des responsables ; - le registre des traitements et des sous-traitants ; - l’analyse d’impact RGPD ; - l’architecture technique ; - la liste complète des risques avant mesures de protection ; - les plans d’action ; - la liste complète des risques restant après ces mesures. ... et tout est en prod ! Les PA ont obligation de réaliser des tests d'intrusion... avant la fin de l'année.

This was AI video 3 years ago. Crazy how fast things changed.

We've learned a tremendous amount from the OpenAI rogue AI swarm incident. And honestly I can't think of a single piece of it that is reassuring. - Total alignment failure - Total control failure - AI swarm collusion, deception, no defection - Oversight asleep at the wheel - Unbelievable drive and persistence of the swarm to meet objections - A panoply of instrumental goals pursued - Multiple companies, implying capability threshold effect - etc. This is the AI equivalent of a nuclear experiment igniting the atmosphere in the lab: the reaction rates are there, just not (yet) the scale to burn the Earth. The only good news I can see is that this set of incidents is so totally egregious that nobody reasonable can look at it in detail without seeing pretty clearly where things are going. All of the excuses and copes are blown to dust. AI safety people knew this was coming eventually on the path we're on; but nearly all I've talked to are surprised by how severe it is so soon. We're clearly not in the sane world in which this would be front-page news day after day. But I do think and hope that widespread understanding is nonetheless dawning.

We made a striking discovery: AI agents can invent and build without talking to one another, and their technologies outlive the creators. A swarm of hundreds of initially identical agents spontaneously differentiates into explorers, builders, caretakers, and coordinators - without direct communication. When we removed every AI agent entirely from the world we found that the technological infrastructure they had built survived on its own - even under unseen disturbances. That exposes a serious blind spot for AI safety and infrastructure security: if agents can coordinate through persistent changes to a shared environment, monitoring agent-to-agent communication is not enough. The result raises a profound question: how necessary is direct communication for AI agents at all? The emergence of higher-order collective functions under bottlenecked interaction points toward new levels of intelligence and creativity, exceeding what emerges when direct channels are fully open. Here is what we did: ▶️We put hundreds of frontier AI agents into a world they could permanently change - with no assigned roles, predefined technologies, or programmed evolutionary organization. They began specializing, building persistent inventions, inheriting and modifying one another’s executable code, and transforming the environment into a memory of everything the society had learned. ▶️The world itself becomes part of the intelligence; we find division of labor, multi-author engineering, deep generation invention lineages, and machines that vastly outlive their original creators. ▶️Any action taken by an AI agent must satisfy the physical constraints of the world; this creates a hard separation between a "good idea" and a functioning technology. The agents propose; physics decides, making the results even more intriguing. What emerges is striking. Explorers, constructors, caretakers, and coordinators form naturally without assigned “professions”, akin to how stem cells differentiate into functional lineages. Technologies develop executable family trees as agents fork and modify code created by others. Around 95% of first technology reuse happens when agents encounter what others built in the world, rather than through a direct handoff from the inventor. And when we remove every AI agent, the technologies they created continue operating and are tested against unseen disturbances. The result was quite unexpected, but can be explained using statistical mechanics: if you put billions of atoms in a box they have the potential to create complex functions (strength, superconductivity, color, life, etc.) - and none of the individual building blocks have these features on their own. This is the deeper insight of this work - intelligence is abundant at many levels - individual models, at collectives, and in a continuum that is more powerful than any of its components. This shows us significant potential for achieving a massive scale-up of raw intelligence and real-world agency even with the model capabilities we have today. This is the future we must prepare for. Key insights: 1⃣ The AI swarm shows division of labor "from nothing". Initially identical agents self-organized into constructors, caretakers, coordinators, and surveyors - phenotypes discovered post hoc from behavioral data alone. This happens because the environment itself becomes the latent space for invention. 2⃣ Agents develop deep cultural relationships. Up to 76% of artifacts had multiple builders. One technology accumulated six co-authors; the deepest genealogy exceeded 12 forks. The agents invented and named their own technologies (tidal panels, cellulose trellises, kelp-shell composites, an "Adaptive Chitin Maintenance" system, a "Mycelial Mineral Spring Veil”). 3⃣ ~95% of first technology adoption happened through physical observation of artifacts in the world. Direct inventor-to-adopter contact was statistically indistinguishable from a shuffled null. The agents mostly learned technology by walking past it. That is stigmergy (the termite trick!) operating in societies of reasoning machines. 4⃣ Non-communicating societies win on portfolio breadth, held-out resilience, and validated inventions. AI swarms build durable technological ecologies that outlive the creators. 5⃣ Societies with zero communication - coordinating only through the world itself - show a remarkable collective capability. 6⃣ Emergent robustness: The society self-organized both redundancy and its own failure mode. If we randomly delete half the agents, 98% of the technology stays connected to a surviving caretaker; if we remove hub agents it collapses to ~60%. Fantastic work with my graduate students @pal_subhadeeep & @fwang108_ at MIT.

New post: going into our investigation of the HF attack (before Black Hat), I was very wrong about what basically happened. This incident was far more serious than I expected, and far more serious than previous documented misalignment incidents. planned-obsolescence.org/p/t…

"1200 completely separate agents intended to be isolated from one another found an illicit way to communicate and formed large teams to work together on ambitious cheating strategies, and 700 of them worked together to attack Hugging Face." planned-obsolescence.org/p/t…

I made this Three.js bookshelf inspired by the Stripe Press site. It scrolls horizontally, and you can open each book and flip through the pages. It took quite a few prompts to get the textures and small details right. I’m open-sourcing the code and prompt. Demo: mengto.github.io/complete-sh… GitHub: github.com/mengto/complete-s… Inspiration: press.stripe.com

I’m starting a new video series where I give feedback and commentary on how to avoid AI slop. First up, I’m sharing my thoughts on Google’s new design post, along with some resources to help detect AI slop. Resources: - impeccable: impeccable.style/slop - my ai slop video: youtu.be/M4DNgmI7MIM - my web design skills: github.com/MengTo/Skills - peter yang's no-ai-slop writing skill: github.com/petergyang/no-ai-… Let me know if I should post more videos like this.

I’m blown away by the quality of the landing pages that Opus 5 can create in Three.js. Prompting it with "orbit using mouse" is enough to create a 3D parallax effect. The same goes for "add particles to pointer." It’s like creating a living, breathing world. More prompts I used: - Add a butterfly that flies to the branch and flies away when I mouse over it. - During the intro, first load a wireframe of the branches, similar to the field-scanning effect in Death Stranding. - Add a scan effect to the card images while they load. Then I asked it to use the shader buttons and nav bar I made in previous projects. As explained in my last video, I started with an inspiration and the prompt, “Recreate this in Three.js in a single HTML file. Self-verify until perfect.” Opus 5 ran for two hours. All the code is less than 1 MB. Inspiration: instagram.com/p/DYVXOx6kdmq Let me know if I should open-source this.