Our strongest frontier model, Gemini 4 Carbon, is now available.
Built for deeply complex, long-horizon workflows that push a frontier model’s intelligence, Carbon delivers frontier performance in agentic workflows across complex knowledge work, software engineering and web development, cybersecurity, and scientific reasoning.
While Gemini 4 Argon remains the ideal model for the majority of software engineering and knowledge work, Carbon steps in for your hardest tasks that push a model to its limits — from complex, multi-step debugging to open-ended scientific research.
Available starting today:
— Consumers: use it in the @GeminiApp and AI Mode in Search for Google AI Ultra subscribers
— Developers and Enterprises: build via the API in @GoogleAIStudio and @AndroidStudio, Google @Antigravity for Google AI Ultra subscribers, and Gemini for business
— Scientists and Researchers: apply for a chance to get 6 months free via @Antigravity
— Cybersecurity Defenders: apply to join the Fairwind Program for access to Carbon for cyber
Watch how Carbon brings reasoning and coding together, turning avian research into an interactive 3D simulation of a songbird’s vocal organ that synthesizes the sounds of six bird species live.
Introducing Mini 2, the next-generation portable Starlink. Get up to 4.5 hours of connectivity on a single charge with the quick-swap battery, or plug in with USB-C.
With built-in Wi-Fi, all you need is a clear view of the sky to get online in 60 seconds.
Preorders are now open by invitation in the US. Be the first to know when it's available → starlink.com/mini-2
I LAUNCHED A GROK BOT TERMINAL OVERNIGHT, GAVE IT $133.70 AND STOPPED ASKING AI WHAT IT WOULD DO. BY MORNING IT HAD $7,351.00 AND SIX AGENTS LOGGING EVERY TRADE
For a year I kept asking agents what they would trade and then just watched the replies scroll by. Last weekend I finally built something that trades on its own, inside Grok Bot, without writing a single line of code myself.
Six agents on one screen, one job each:
> scanner catches new pairs and volume spikes, flags only, never trades
> rug check kills suspicious tokens before anything gets executed
> sizer sizes every position from account equity, never a fixed ticket
> executor opens and closes trades and tracks slippage, capped at 2%
> logger writes down the reason behind every single decision
> equity updates the balance and the curve on every close
I went to sleep at 23:00 with $133.70 on the account and the dashboard left open. At 07:00 the balance said $7,351.00. 64 trades, 43 wins, 21 losses, and every one of them had a reason written next to it in the log.
That's the part I didn't expect. Not that agents made money overnight. It's that I stopped being the guy asking "would you long this?" and became the guy reading the log of what already happened.
Starting balance: $133.70
Current balance: $7,351.00
Grok Bot built the terminal, the agent setup and the dashboard. I just described what I wanted and it went live.
The full build is in the post below ↓
1⃣ Comment "Grok_Bot"
2⃣ Like and Repost
3⃣ Follow me @Corki_0x
I'll DM you the complete setup.
Hazwan S. retweeted
Your agents can now coordinate with your teammates’ agents.
Projects has been rebuilt for multiplayer: every teammate and their agents can now work together in one unified workspace.
Hazwan S. retweeted
Updated the survival skills app with more lessons!
Learn how to navigate with or without a compass, find your way at night.
Treat wounds, tie knots, prevent hypothermia, find water, and more.
Available in 21 languages
backcountry-101.vercel.app
Hazwan S. retweeted
I got a 1.9B model to watch a restaurant camera and tell staff which tables need cleaning 🍽️
It runs on my laptop. No cloud, and it generates zero tokens.
Built over a weekend with @liquidai d1-2B
Hazwan S. retweeted
Claude Fable 5.2 is now available in Cursor!
At max effort, it's the top model on CursorBench at 60%.
Fable 5.2 scores 56.9% on CursorBench at default effort, up from 49.2% for Fable 5.1.
Hazwan S. retweeted
𝗕𝗿𝗲𝘄𝗲𝗿𝘆 𝗳𝗶𝘁-𝗼𝘂𝘁. We gave Plamo 5 marked-up plans, annotated in metres and feet. One wrong conversion, and the equipment doesn’t fit the room.
Plamo reconciled the units and reviewed the layout item by item before any more modeling. Swipe through the building, then to the last slide for the coordination table.
Save this for your next fit-out review, and start designing with Plamo today: plamo.ai.
Hazwan S. retweeted
Put a Gemma-based model in a robot body. First thing it does: scan the room, spot the giant TV, and start a movie break.
Awesome to see @NVIDIAAI build on Gemma 4 E4B, fitting live conversation, tool calls, animations, and a vision model onto a single 8GB Jetson!
Hazwan S. retweeted
All right, Blender + Claude + Seedance is just really really fun. This took me less than 30 minutes to make... And I didn't know anything about Blender.
Can't wait to try it on a short movie.
Hazwan S. retweeted
AI Studio now lets you build rich apps with Google Maps with just a simple prompt! 🚀
Our batteries included experience to build apps with the Google ecosystem continues to get easier :)
GrokBot engineers just released a 11-page PDF on GrokBot token usage and how to cut it from 100% to 20%:
why GrokBot burns your weekly pool, what eats the most and how to cut usage from 100% to 20% in my 11-page research:
step 1 → usage is one equation: wake-ups × context × steps × fan-out. The factors multiply, so four small cuts beat one big one
step 2 → it's one pool. Grok chat, GrokBot and Cursor drain the same weekly %. No per-bot breakdown, no warning before 100%
step 3 → wake-ups burn first. routine every 15 min runs 672 times a week, even when nothing changed. Test runs count too
step 4 → swap schedules for triggers. Slack mention, email, GitHub event, webhook. Same job, 672 runs → 15
step 5 → every reply re-reads the whole thread. GrokBot on one job for 6 weeks carries all of it into every run. Retire it with 300-word handoff note
step 6 → trim the profile and memory. They ride along on every step, thousands of times a week
step 7 → give every task finish line and a stop rule. Fails twice? Stop and ask. Retry loops are the worst leak
step 8 → apps over clicking. Open, screenshot, scroll, click is 7 steps. A connected app is 1 call
step 9 → one owner per job. 4 GrokBots in a group chat = up to 4 billed replies per question. Cap bot-to-bot at 2 messages
step 10 → at 80% midweek switch to low-usage mode. Pause routines, final results only, set a spend cap before 100%
the result: same jobs, 2,000 units → 400. Wake-ups alone cut the week in half. pool didn't get bigger. Your GrokBots got leaner.
this 11-page PDF teaches you how to never hit your GrokBot limit again
save now, then read how to save GrokBot tokens in the article below
your GrokBot hit 100% with 4 days left
no reset coming. Elon Musk called it "unsustainably" expensive
you're not using GrokBot too much. you're using it wrong
in this article, I show 10 settings and copy-paste prompts that make your pool last all week:
Readers added context they thought people might want to know
The claim that GrokBot engineers released this 11-page PDF is false. No such official document exists from xAI or the GrokBot team; it is the poster's own user research.
x.com/grok/status/21…
docs.x.ai/grok-bot/how-g…
x.ai/bot
x.com/elonmusk/statu…
Hazwan S. retweeted
Replying to @bot
@bot can now connect to private MCP servers on your tailnet without exposing them to the public internet.
Powered by Tailscale OAuth apps, now GA, so you can build secure agent integrations, too.
Read more here: buff.ly/U1J9By1
Hazwan S. retweeted
you can now build with google maps platform in AI Studio 🗺️📍
go from idea to app powered by @googlemaps in minutes!
what’s new:
- 1-click setup
- create maps apps just by describing your idea, gemini does the rest
- build with map tiles, places, weather, routing, 3D buildings, street view, and more
try it out today: ai.studio/build
Hazwan S. retweeted
You can now connect multiple local computers to the same dot and choose which one it uses for a task.
It can build and test your app on a dedicated machine while you keep coding on your laptop.
Hazwan S. retweeted
RustGIS v0.6.0 is here!
RustGIS is a fast, free and open-source GIS in pure Rust for mapping, analyzing and sharing geospatial data. It runs natively on Windows, macOS and Linux, and in your web browser. Your data stays on your machine unless you send it to an AI model you choose, and every action is a command you can script.
This release runs your tools in the background with progress and Cancel, exports layouts as vector PDF and SVG, and adds diagrams, label callouts and new polygon editing tools. It brings together more than 250 pull requests from 29 contributors, 8 of them contributing for the first time. Thank you to everyone who helped!
What's new in v0.6.0
- Background work: tools, exports and large reads run off the main thread with progress and Cancel, many on several cores, plus batch runs.
- Publication output: layouts as vector PDF and SVG, multi-column legends, and maps exported as georeferenced GeoTIFFs.
- Cartography: pie and bar diagrams, label callouts, Move Label, and MapLibre hillshade and colour relief.
- Editing: Cut Polygon, Reshape and Trace, with Check Geometry and Repair Geometry to fix invalid shapes.
- More data: CityJSON city models, ArcGIS protocol buffer pages and cloud-optimised GeoTIFFs read in place.
- Workspace: a History pane of tool runs, custom themes, UTM and MGRS readouts, and Greek and Lao.
Try it out
- Launch the web app: web.rustgis.app
- Download: downloads.rustgis.app
- Documentation: rustgis.app
- GitHub: github.com/opengeos/RustGIS
- Release notes: github.com/opengeos/RustGIS/…
#GIS #Rust #OpenSource #Geospatial #Cartography #RustGIS
Hazwan S. retweeted
Building blocks for interactive UI, ready to use as a library, skill or mcp with Claude Code, Cursor or Codex.
44+ transitions, 30+ interactive components
transitions.dev
Hazwan S. retweeted
i still don't understand why everyone is NOT fixing their Grok Bot usage
it will probably make your bots last the whole week & definitely change your life...
i wrote the best article on exactly what your bots need to stop burning your limit
simply paste this article into grok bot, it will find what's eating your limit and fix it
this is the EXACT setup i ran on my bots, and my limit hasn't run out mid-week since:
Hazwan S. retweeted
Keeping a high quality bar + performances + doing the whole world in nice cute 3D is such a challenging problem.
I'm very close to being able to scale to the whole world technically.
All of those already available on Anywhere:
apps.apple.com/us/app/anywhe…






























