Adam Killam

Gemini advice for next 10 years.

Agents are your new customers. Introducing Intents by @tryscope_app tryscope.app Now you can see why they discover your product, and why they leave. See what agents came to do, where they get stuck, and what's missing from your docs, content, and features. We draft the fix. You hit publish. Then track the impact.

A ‘hello world’ milestone for Project Suncatcher 🚀 After launching our prototype satellite a few weeks ago, we’ve now successfully sent a command up to orbit, booted up the TPU, ran the prompt, and received our first complete response back on earth - a haiku!

The most important book right now for understanding company building in the age of AI is about an agent... a hollywood one... Ari Emanuel's book ('roll the calls') -- there are TWO key lessons from it (one simple, one a bit more nuanced but even more important...) The two: First - the easy but important point... it is all sales right now as software commodifies, and the real lesson of first half of the book is simple -- as software commodifies and everything is just sales, if you get a 'no' call the next day... over and over and over... be relentless and keep asking til you get the answer you want. This mentality has historically not been in software / tech CEO DNA, but as the leverage of being 'unique genius' gets destroyed by arms-dealing AI / commodification of expertise (and certainly software) -- the winners are storytellers and salespeople. Second - the more intellectual but even more important point - WHAT AGENTING REALLY IS? Why does anyone pay an agent... for trust / the relationship. What do agents do? Whatever the clients need / everything - not one thing. Why this matters? it totally changes business building. In a world where tech is 'hard' being a leader is figuring something out about the universe, building an advantage around that competency, and doubling down over and over on your 'advantage' in providing something specific. No more... whether for consumers or for companies... if you are a 'business' in an agentic world, your asset it TRUST of your client, and you do as much for them as you can / anything (also send them random gifts). Your job is to get in the door, become trusted, and then pick up more and more business doing whatever the customer needs. This means the mission is SERVICE not PRODUCT or MISSION. The asset is the 'relationship' ... and the winners do everything they can for their clients. If you are building a modern business right now - this reality changes everything from how tech companies were built over the last 20 years or so... mission is less important, top-down vision is less important, instead what matters is sales, clients, and being flexible with 'free to develop' software and services to SOLVE problems for your actual specific customers. It this GOOD for most (startup) business... probably not. Agents get fired all the time, there is less enterprise value in what they build, etc... also per Ari's book, in the agenting world, scale is everything -- but from a technological deterministic perspecitve... this is where we are going, and founders that understand that (relenteless sales, how do i get in the door with something and build trust, and then what I 'do' is everything ... will be the RELATIVE winners)...

For fifty eight years, my boss in Tokyo believed that beef should be sliced as thin as paper. In Japan, we cook thin meat for four seconds in boiling broth. That is our way. Then last month, the YouTube algorithm showed Director Sato a video of a man in Texas smoking a brisket. It changed his life. On Monday morning, he called me to his desk. His eyes were bloodshot. "Have you seen this?" he whispered, turning his monitor toward me. "The American pitmaster does not sleep for sixteen hours. He sits by the fire in the dark. He guards the meat like a castle." I said, "Sir, it is just a barbecue." He slammed his hand on the desk. "This is not cooking," he said. "This is a vigil." Last Saturday was our annual company picnic by the river in Kanagawa. Normally, we grill thin meat and yakisoba for two hours and go home at 2:00 PM. I arrived at 9:30 AM to help set up. Director Sato was already there. I want to be accurate here: he had been sitting in a folding chair since 10:00 PM on Friday night. He was wearing a towel around his head and staring at a massive black barrel smoker he had imported from Oklahoma. Inside I said: THIS MAN HAS ABANDONED CORPORATE LIFE FOR THE SMOKE. When we tried to help him, he held up a pair of metal tongs and clicked them twice. "Step back," he said calmly. "The stall has begun. Nobody opens the lid." Twelve Japanese accountants stood in complete silence around a barrel for four more hours. Nobody dared to eat a single noodle until he gave the signal. At 1:45 PM, he sliced the meat. It fell apart on the cutting board. My boss looked at the pink smoke ring, nodded once, and wiped a single tear from his cheek. He has already submitted an official budget request for a trailer smoker for next fiscal year.

Introducing Gemini 4 Carbon, it's a big jump over Argon which we released just 2 weeks ago. Gemini 4 Carbon is SOTA on cyber, long horizon tasks, ranks #1 on Artificial Analysis V5, significantly improves on usability in domains like coding, and is more aligned.

It’s kinda crazy to me that we melt sand down, clean it to 99.99999% purity, literal wizards have to draw it slowly into an ingot, we cut it, inscribe a hundred billion little on/off switches on it, and when you run electric current through it you can do basic arithmetic, which when you do enough of it can do any math on the planet, and which describes reality well enough that we can teach it the ontological meaning of words and their distance from each other on a high-dimensional space, and if you ask, it can tell you what word comes next and that apparently is powerful enough that if you do it a hundred billion times it’s the smartest thing to ever exist oh yeah and we use this primarily for email

Happy to announce that we are live and accepting merchants 🫡. Connect your e-commerce or online store once and Kifly allows you to sell your products through AI agents in many channels. Nike and now Walmart are building their own in house infrastructure to serve agents. Kifly is built for the small to medium size merchant who can't. 1 - On your existing store, you get a sales agent that understand your products, reviews, policies and can help customer checkout. 2 - Connect your own agent to Instagram and other places customers reach out to you. It monitors your DMs, sends product links, creates a cart and hands the customer a checkout link to pay. Offering a seamless experience to shoppers. 3 - You get your own AI agent facing store through an MCP that Muse, ChatGPT and Claude can access to search, cart and get a checkout/payment link to pay on behalf of buyers. Kifly's native payment rail is powered by @stripe, and compliant with UCP. The future of shopping is here, is your store ready?

Ok, I guess now is a good time to announce my return to X... Burp! Enjoy.

Introducing Flint, @Recursive_SI’s first model: a fast local coding model built with our automated research system, the Eureka Machine. • Best-in-class fast local coding • Up to 2x faster than comparable models: 125+ tokens/sec on a MacBook Pro, 150+ on RTX Spark • Built from scratch in under 4 months Private, unmetered intelligence on your own machine. Open weights available now on @huggingface, run it locally on your favorite framework or try it for free on @opencode or @pidotdev.

1/4 Today, we’re releasing Mercury 3. Mercury 3 is by far our most powerful model and the most intelligent diffusion LLM ever released. It outperforms GPT-6 Luna and Haiku 5.5 on quality while running at 1000+ tok/sec.

Spent years watching models ace a single image and completely whiff on anything that moves or changes angle. Today we’re fixing that. Try Glimpse now, our first visual thinking model with competitive frontier performance compared to Astra on key public visual reasoning benchmarks.

we're dropping something heavy tomorrow @every it's the 8k word definitive history of Codex written from deep interviews with the team that did it including @gdb @thsottiaux @ajambrosino @embirico @mia_glaese @edbayes and more it'll also be agent-native so you and your agent can explore the full text of all the interviews in addition to reading the story :)

after switching our approach to building a "FirmOS", I'm increasingly confident that the future of agent collaboration within organizational settings is going to be a "company operating system" or CompanyOS, designed for agent and human collaboration instead of a "company brain" that holds everything, the OS is for storing rules, processes, decisions made, which tools to use and how. it can mirror some data, but the source of truth is often best held in your existing tools (CRM, etc.). it's continuously updated and used by your agents i think of this as a "flipped playbook" where instead of writing down the company rules first and following them (which is kind of like waterfall development), you're constantly updating the playbook as you decide or learn new things day-to-day running the business