标签：#AI

On the AI video ad platform I work on, every scene goes through the same painful loop: write a prompt, send it to an AI video model provider, wait two minutes, open the result, squint at the frame, and decide what went wrong. Camera too wide. Product missing from the hero shot. Color palette drifted warm when the brand brief says cool neutrals. Avatar looks like a different person than scene three. That loop was manual, slow, and expensive. Each regeneration burns GPU credits. Operators were becoming prompt engineers by accident — and still missing subtle failures until stitch time, when fixing scene four means re-rendering everything downstream. The insight behind vision-in-the-loop prompt authoring is simple: the model that wrote the prompt can also look at its own output and rewrite the prompt with surgical fixes. Not a full replan — a per-scene correction grounded in the actual generated frame, not the operator's memory of what they hoped would appear. The manual loop we were trying to kill Before this work shipped, the swipe iteration flow looked like this: Plan — Claude generates a scene-by-scene script with visual prompts Generate — each scene renders independently through an AI video model provider Review — operator opens the portal, compares frames to the reference ad Rewrite — operator edits prompts in a text field, often guessing at what the model misread Regenerate — repeat until acceptable or budget exhausted Steps three and four are where throughput dies. An experienced operator can spot "product not visible" in three seconds, but translating that into prompt language — "medium close-up, product centered in lower third, shallow depth of field" — takes another minute per scene. Multiply by twelve scenes and three swipe iterations, and a single ad creative consumes an hour of human attention that should be spent on brand strategy, not frame inspection. The generated frame is ground truth. The original prompt is a hypothesis. Vision-in-the-loop closes the

Meanwhile, paid accounts can configure prompts to trigger when things happen in Gmail, Slack or GitHub.

Record-breaking robot races are less substantial than household chore challenges.

CD Projekt Red is remastering The Witcher 3: Wild Hunt, the hit RPG that first launched in 2015, and the updated game will be available starting September 29th. The Witcher 3: Wild Hunt - Remastered will bring improvements to visuals and combat, a "reimagined" skill tree, and more, and it also includes the first two […]

The company's new fundraising total now stands at $232 million.

El chatbot va último: cómo ordeno las cinco automatizaciones que más me piden Tengo 16 flujos en producción para pymes y la lista de pedidos se repite casi siempre igual. Lo que no se repite es cuál conviene hacer primero. La discusión habitual las ordena por dificultad de construcción, y esa es la métrica equivocada. Construir es la parte barata: el modelo escribe la mayor parte. Lo que se paga después es el mantenimiento, y ahí el orden se da vuelta. Van las cinco, ordenadas por lo que cuesta sostenerlas, de la peor a la mejor. 5. Responder consultas frecuentes La que todos piden primero y la que más mantenimiento tiene. Parece contenida: son las mismas veinte preguntas. No lo es, porque el contexto que necesita se mueve todo el tiempo. Cambia el catálogo, cambian los precios, cambia el horario en verano. Meta cambia requisitos de la API. El modelo se actualiza y el mismo prompt deja de comportarse igual. Y sobre todo: es la única de las cinco donde el error lo ve el cliente . Un bot que inventa un precio no genera un ticket interno, genera un reclamo. Si igual va primera —y a veces va, porque es la que se ve—, presupuestala con el abono adentro desde el día uno. 4. Turnos y reservas La más engañosa de la lista. Tomar un turno es trivial; el problema es todo lo demás. Cancelaciones, reprogramaciones, dos personas pidiendo el mismo horario con cuatro segundos de diferencia, el turno que se cargó a mano en el sistema y el bot no vio. Es estado compartido con escritura concurrente , que es un problema viejo y conocido, disfrazado de chatbot. Si la agenda vive en un sistema con API decente, baja bastante. Si vive en un Google Calendar que además tocan tres personas a mano, no la subestimes. 3. Mover datos entre sistemas La que más valor devuelve y la que menos depende de vos. El trabajo real casi nunca es la transformación de los datos: es el sistema del otro lado. Y en pymes ese sistema suele ser uno de gestión local, sin API pública, sin documentación, y con un prov

Short answer: use an emailed, single-use reset link as the default for most SaaS login recovery, and add SMS OTP only where users may genuinely lack email access or the product already maintains verified phone numbers. Email is usually the simpler system because the login identifier, recovery destination, and support workflow can remain in one channel. SMS can shorten the interaction, but it adds phone-number lifecycle, message segmentation, regional consent, and delivery-state work. “Cheaper” depends on your traffic and failure rates, so model completed recoveries rather than message sends. This is a recovery decision, not a notification preference. The goal is to return the right person to an account without turning a delayed message, an expired credential, or a recycled phone number into an account takeover or a support queue. I've worked around enough spam filtering, rate limiting, and OTP delivery gaps to treat the channel as one component of that system — never as the system itself. What should a SaaS password recovery flow use: email API or SMS OTP? Start with the account data you can already trust. If every user signs in with an email address and changing that address is a controlled operation, an email reset link creates the smaller data surface. The service generates a high-entropy, single-use token, stores only a protected representation of it, sends a link, and accepts that token once before a short expiry. The browser then moves the user into a password-change session. An SMS OTP flow looks compact on screen, yet the backend has more questions to answer. Was the phone number verified recently? Can the user update it without being signed in? How are country codes normalized? What happens when a number is reassigned? Does the support team have a safe path for a person who lost the device? A six-digit form doesn't make those policy decisions disappear. So the default is straightforward. Choose email first when email is the stable account identifier and rec

You do not have a camera. You have a machine that dreams a short motion out of a single still image, and it dreams badly the moment you ask it for something the still does not already contain. I learned this across a 10-episode series, and every rule below was paid for in failed generations. None of it is theory. The medium's real physics A real camera moves through a space that exists whether or not you point at it. The model has no space. It has one flat image and a statistical guess about what "zoom out" tends to look like in its training data. When the frame widens, the model is not revealing more of a room that was always there. It is inventing pixels to fill the new area, drawn from everything it has ever seen. That single fact reorganizes everything you know about directing: There is no coverage. Every "angle" is a separate generation from a separate still. Continuity is not captured; it is engineered, frame by frame. Nothing survives the cut for free. The model does not know that shot 12 and shot 13 are the same character in the same room. Anything you want to persist (damage state, light, color) must be re-declared or re-anchored every single time. The model abhors an empty frame. Its deepest reflex is to resolve ambiguity: a silhouette becomes a face, fog becomes a mountain range, a clean retro interior grows drips and cobwebs because "analog" reads as "abandoned". Spawn pressure is constant. Background figures flicker into existence in any populated-looking scene. Every motion prompt in my pipeline ends with an anti-spawn guard: "Do not add extra characters. Keep everything as pictured." Drop that guard and the figures come back. A widening or traveling frame is an invitation for the model to hallucinate. Direct this camera and you are not choosing what to show. You are choosing what to withhold from its imagination. The classical grammar, re-pointed If you carry film vocabulary, it all still applies. The mechanism just changes completely. Classical tool

Anthropic is giving Claude a shared memory across chat and Cowork, so users no longer have to repeatedly brief the AI on projects, preferences, and other context.

It’s time for another Geoff Keighley-hosted blast of gaming news. Gamescom’s two-hour Opening Night Live event is scheduled to kick off on Tuesday at 2PM ET, with a preshow starting 30 minutes before. Thanks to teases on social media, we already have some idea of what to expect, including looks at Final Fantasy VII Revelation, […]

Anthropic is giving users more control over what memories Claude saves.

You've written it a hundred times. In your CLAUDE.md , in your system prompt, in ALL CAPS: NEVER put "use client" at the page level. NEVER commit @ts-ignore without a reason. And your agent does it anyway. Not always — that would almost be easier to deal with. It follows the rule for the first 50k tokens, then quietly stops. Or Sonnet follows it and Haiku doesn't. Or it follows nine rules and forgets the tenth. Here's the thing I finally accepted: a rule in a prompt is a request. The model can decline it. So I stopped asking, and started enforcing. TL;DR Prompt adherence is probabilistic. It degrades with context length and with model size. But half of my coding rules never needed a model at all — they're grep-able. Claude Code hooks + exit 2 turn those rules into a deterministic reviewer that runs after every single edit , costs zero tokens when nothing is wrong , and fires at 100% regardless of which model wrote the code. Once the mechanical rules are enforced from below, you can safely downgrade the model doing the typing. That's the real payoff. Everything below ships in ccteams v0.3.0 , but the pattern takes 30 minutes to build yourself. Two kinds of rules Some background in three lines: I run Claude Code with orchestrated agent teams — a builder writes code, a reviewer verifies it, and both get a stack-specific "playbook" of rules distilled from the mistakes mid-tier models actually make. It works well. I wrote about the prompt-engineering side of it before. But rereading my playbooks, I noticed the rules split cleanly into two categories. Rules that need judgment: Trace the Server/Client boundary by hand. Don't write a fix until you can state the root cause. These need a model. Prompts are the right place for them. Rules that are just string matching: "use client" at the top of app/**/page.tsx → wrong. process.env.SECRET in a client file → wrong. @ts-ignore with no justification → wrong. Why was I asking a language model to remember these? A regex doesn't get

A pioneering AI scientist once predicted computers would replace human radiologists. They haven't.

Lica co-founders are going to work on Gamma's new research team.

Tested on SemiAnalysis’ InferenceX benchmark, Jalapeño registered both more tokens per user and more throughput per kilowatt than the currently available state-of-the art.

The funding, Gatik's largest so far, was led by Qatar Investment Authority and Koch Disruptive Technologies.

OpenAI says its new AI chip, Jalapeño, completes tasks more efficiently and returns responses faster than other AI systems, according to a blog post published on Tuesday. During a briefing with reporters, OpenAI hardware vice president Richard Ho said Jalapeño offers the "best of both worlds" with lower latency and higher throughput, as AI systems […]

We're liveblogging the two-hour showcase and half hour pre-show of Gamescom's ONL today! Come join us, we'll even start a little early for you.

Apple finds respectable performance gains on a new 2nm process, but M5 users shouldn't feel any FOMO.

Folks have been daisy-chaining Macs for AI—this refresh keeps that in mind.