Pokee AI retweeted
I believe models close to today’s frontier capabilities will run entirely on local hardware within two years.
At Valley101’s Alignment 2026, I shared why—and what we’re already building at Pokee AI.
Pokee-Isaac runs a 30B+ parameter model on a standard AI PC with 32GB of RAM:
• 200K-token context
• 1,200+ tokens/sec prefill
• 20+ tokens/sec decode
The shift is driven by two forces: people want more control over their data, and smaller models are catching up to the frontier.
Models keep getting more efficient. Local compute keeps getting stronger. I expect those curves to meet within two years.
Pokee-Isaac is an early look at that future.
#LocalAI #OnDeviceAI #PokeeAI
Pokee AI retweeted
I ran the same task on Claude Code and DeepSeek's new agent harness. One cost $150. The other cost $2.
Today we're launching AgentSky.dev (@agentsky_dev), the "OpenRouter for Agents" — one API → Claude Code, Codex, DeepSeek, Kimi, OpenCode, and every major agent in the cloud.
And Agent Playground on top: race them on your own task, with your real tools (GitHub, Gmail, more), side by side in a browser: time, cost, tokens burnt.
Guess which one was $2.
Pokee AI retweeted
Built an AI cinematic journey with @Pokee_AI.
6 generated scenes. One continuous story. Countless iterations to get the pacing and transitions right.
AI tools are becoming more than generators, they’re becoming storytelling workflows.
Excited to keep exploring what’s possible. 🚀
Pokee AI retweeted
COOL.
'Faith' music video, end to end w/@Pokee_AI's Issac!
Isaac directed: shot plan, continuity, prompts, and the tighter multi-cut edit structure.
Grok @Imagine made the plates.
my Arkhē Lyric Lock timed my lyric sheet.
Assembled locally to the master.
First cut with 10 long slots looped and felt soft.
Isaac’s expanded cut (20 segments, no loops) + inserts is a different film.
• Real director job: concept, continuity, face-late, clay spine — held across the project
• Prompts that Imagine could actually shoot (not vague sludge)
• When long slots looped, Isaac redesigned to more cuts + insert plates — not “make longer AI”
• Cheap API for that brain vs what the look cost in Imagine/time
Original Music and Lyrics by me, Adem Vessell 🫡
AI Digital cover by
Pokee AI retweeted
Everyone thinks 10M context is for reading giant PDFs. Wrong❌
It’s iteration memory.
I just built a playable 3D FPS in one HTML file using Pokee’s new Isaac model 🔥
Shot 1 broke. Shot 3 worked. The 10M window patch its own code in-place without losing the plot.
@Pokee_AI
Releasing Pokee-Isaac 28B — the world’s first real 10M-token context frontier-class agentic model, deployable on a single GPU (starting from RTX 4090 or equivalent).
New proprietary non-decoder-only architecture:
• 93.3% RULER at 10M tokens
• Up to 137K tokens/s prefill on one B200 with 10M-token context
• Leads BFCL v4 and τ³-bench in our evaluation
• Lowest combined attack success rate among evaluated models on DTAP security red-teaming benchmark
Pricing and deployment:
💰 $0.15/M input · $1/M output
🔒 Deploy in your VPC, on-premises, or on-device, with Day-0 support for @vllm_project and @sgl_project
Technical blog: console.pokee.ai/model
Technical report: console.pokee.ai/pokee-isaac…
API: console.pokee.ai
Pokee AI retweeted
I tested Isaac (Pokee) on a real use case.
A music studio: intention → finished track.
Style, vibe, duration..
Two minutes later, the audio is ready.
Isaac handles analysis and planning end-to-end.
10 million tokens of context.
No chunking. No chatbot.
Full track analysis: 1 cent.
The system runs autonomously all the way to display the result on the player.
@threejs @MeshyAI @Pokee_AI
#Pokee #AIAgent #music
Pokee AI retweeted
Already making some cool things with the newly released @Pokee_AI Isaac model. I was tired of having to find tracking emails, so Isaac built me a dynamic dashboard that scans my Gmail and keeps me updated through a daily report delivered by my Hermes Agent.
Pokee AI retweeted
Okay, this is interesting.
Why should teams have to choose between a model small enough to run locally, a massive context window, and real agentic performance?
Pokee-Isaac 28B V0 is built to deliver all three.
10M tokens of usable context.
Strong tool use and multi-step execution.
Deployable in your VPC, on-premises, or on compatible devices.
And it does this with just 28B parameters with $0.15/M input and $1/M output token cost.
That’s the part worth paying attention to: long-context AI doesn’t have to mean a gigantic, cloud-only model.
@Pokee_AI
🤝 Paid partnership
Pokee AI retweeted
It takes less than 10 minutes to make an app out of X over on @Pokee_AI. Which includes the X API for free in its low token costs.
I talk to @AndrewWarner about why this is so powerful.
And it is cheaper now to do the same with your OpenClaw or Hermes or Manus or almost any AI agentic system.
It's how I built alignednews.com/ai.
Why is this a big deal? Well, EVERYONE can now build software. I can show you how to do this in a few minutes. I'm not a programmer, and I built this system just by talking to an AI about what I wanted.
"Build me a website of all the important news from the AI community."
Yeah, it took me two months to really teach my agents how to look at the news and what I wanted, but it really is that easy to build software now.
Everyone can build an app for their communities and teach it what to find that's valuable here on X.
I totally missed the value of X's API changes until @Scobleizer explained them to me.
Now, for PENNIES a day:
1. An agent reads & analyzes my X feed to uncover posts that help my business
2. I created a gorgeous app using @Pokee_AI to visualize X posts (I didn't even know wtf Pokee was before, but it took seconds to build!)
3. I'm adding data from X to the internal apps I'm building with Claude Code
- @Scobleizer is using it to build a news site.
- @elvissun is using it to DM people immediately after they interact with his posts.
- @elonmusk encouraged people to use it with @openclaw
Pokee AI retweeted
I didn’t make it clear. Built the two apps with @Pokee_AI
Replit does a bunch of other stuff.
The more tools we learn the more we are able to build.
Met a founding engineer today from @Replit. Jen Li. We were both judging the @Pokee_AI hackathon.
They have me some credits and I built two apps in 20 minutes using the X API:a weather one which mapped storms being reported in by my climate scientist list and another monitoring my three news lists for information about the Iran war.
Pokee includes X API for its customers automatically.
While doing that my other AI read all your posts: alignednews.com/ai
He tells me about why Replit is hugely important in the AI industry. In other words how you can use it in your life and business.
Try using the X API
Met a founding engineer today from @Replit. Jen Li. We were both judging the @Pokee_AI hackathon.
They have me some credits and I built two apps in 20 minutes using the X API:a weather one which mapped storms being reported in by my climate scientist list and another monitoring my three news lists for information about the Iran war.
Pokee includes X API for its customers automatically.
While doing that my other AI read all your posts: alignednews.com/ai
He tells me about why Replit is hugely important in the AI industry. In other words how you can use it in your life and business.
Pokee AI retweeted
OpenClaw has proven that local AI assistants have product-market fit. But the big issue with them has been security.
The team at @Pokee_AI is fixing it with PokeeClaw: works like OpenClaw, but with in a secure sandbox architecture with isolated environments, approval workflows, role-based access control, and audit trails built in. Also lower token usage, which is always nice.
Openclaw is goated, but here is an alternative that takes literally zero setup.
No configuration.
No Mac mini.
No security scares.
I got access to the PokeeClaw waitlist, and it's pretty awesome how fast you can go from nothing to a working agent.
It literally takes 0 seconds.
I recorded a video showing how it works.
If you're new to Pokee and therefore PokeeClaw, check out this article on getting started! We've just launched today, and I can't wait to see what you will all build :).
Pokee AI retweeted
Jensen Huang: "every company needs an OpenClaw strategy."
Pokee AI retweeted
The AI talent market just shifted dramatically.
In the past few weeks alone:
→ Block laid off ~40% of its workforce
→ Meta is cutting ~20% of roles
→ Oracle let go of 45,000 people
That means some of the strongest researchers, engineers, and builders in tech are suddenly on the market — people who were doing exceptional work just days ago.
We'd be foolish not to try to recruit them.
At Pokee AI, we're building the next generation of AI agent infrastructure powered by reinforcement learning — helping enterprises deploy AI systems that stay accurate and reliable across long-horizon, multi-step workflows, even in complex and regulated environments.
Backed by Point72 Ventures, Qualcomm Ventures, Samsung NEXT, and many others for over 12M USD seed round, we're pushing the frontier of RL, LLMs, and production AI systems. The problems are hard, the team is small, and the bar is high.
We're currently hiring for:
🔬 RL AI Research Scientist
🧪 RL AI Research Intern
⚙️ AI Infrastructure Engineer
📈 GTM Specialist
🌍 Remote (US / Singapore)
If you're someone who wants to work on problems that matter at the intersection of RL and production AI — we want to talk. DM me or share this with someone who'd be a great fit.
Apply at pokee.ai/careers
Pokee AI retweeted
Happy New Year! The best way to start a year is to hire the best people for our grand mission! Please help repost to your network!
At Pokee AI, we're building efficient foundation agents that scale infinitely in workflows and tool calls. As we scale in 2026, we're looking for exceptional talent to join our team.
We're hiring:
🔹 Senior Systems/Backend Engineer
🔹 AI Engineer
🔹 Full-Stack Software Engineer
All positions are remote (US/Singapore preferred). If you're passionate about pushing the boundaries of AI and want to make a real impact, we'd love to hear from you.
Check out our open roles here: pokee.ai/careers?reset=1
Don't see a perfect fit? We're always looking for exceptional talent—reach out at [email protected].
Here's to building something extraordinary together in 2026! 🚀
Pokee AI retweeted
We’re kicking off something new for the Pokee community!🪴
Join the newly launched Pokee AI Community — here on X or on Discord — and share your favorite Pokee workflow (how you’re using Pokee for research, automation, or creative work).
Next Monday, we’ll go live to announce one community member, drawn on stream live at random, who will receive one free month of Pokee Lite!
To enter:
1️⃣ Join the Pokee community on X and/or Discord
2️⃣ Share your workflow under this post
3️⃣ Tune in Monday for the live drawing!
We can’t wait to see what you’re building with Pokee!
Pokee AI retweeted
🚀 Introducing SGLang Diffusion — bringing SGLang’s high-performance serving to diffusion models.
⚡️ Up to 5.9× faster inference
🧩 Supports major open-source models: Wan, Hunyuan, Qwen-Image, Qwen-Image-Edit, Flux
🧰 Easy to use via OpenAI-compatible API, CLI & Python API
Built with FastVideo to power the full diffusion ecosystem, and special thanks to @NVIDIAAIDev and @VoltagePark for their compute support!
⬇️Read more in the thread:






























