IF you're not using Ovis with herding your sleeping. I'm going to post a video tomorrow on how to set it up. Why you would want to use it.

Had GPT 6.1 Sol try to solve a problem for 7 hours. Grok 4.7 solved in 20 minutes. GPT 6.1 Sol never did. For those saying Grok can't perform it is still my go-to.

I delivered a website to a customer this week, and they will present it to their leadership for review next week. We'll see how it turns out. Overall, their focus group feedback was overwhelmingly positive.

Can we get a $400 plan? I want to ensure I never run out of Grok Bot or Cursor.

Elon is a savage been loving his posts over the last 24 hours

Hit my first 100% on Grok Bot and it is a generous amount

Grok Bot's work is pretty incredible; the ability to loop and hand off tasks is unmatched. It ships a full software stack, fully tested end-to-end, on different platforms within 30 minutes to an hour. Its ability to catch bugs, send them back, and keep testing saves me an incredible amount of time. If you're still locked into TUIs only, you are heavily crippling yourself.

Wanting to test the refuseless endpoint, does anyone have ideas on what to test it on? I am sure some of the cyber peeps would know.

BLUF: Not worth using based on a few test runs. Below is the repo; it gained traction on GitHub but fell short in practice. Headroom is an open-source compression layer that sits between an AI agent and its model, shrinking tool outputs, logs, RAG chunks, and code before they reach the context window. It runs locally and stays reversible. Agents working on code see a 20% token cut with the same final answer. It ships as a Python library, a proxy, or an MCP server. Anyone running an agent that burns tokens on boilerplate tool responses would use it to cut cost and latency. You can run it as a transparent proxy in front of an existing agent, which needs no code changes. Say the agent pulls a huge log dump from a failing build. Compression shrinks that dump to a small fraction of its original size, and the FATAL line stays intact byte for byte. That still hands the model the error it needs to fix. File reads and RAG retrievals get the same trimming in this setup. github.com/headroomlabs-ai/h…

Usage on new plans is mental... So good

For my B2B SaaS Founders on here

A router receives a packet with a valid destination IP but drops it even though it has a matching route. What could cause this, based on the choices below? A. An outbound ACL B. A DNS cache C. An STP priority D. A DHCP lease Drop your answer 👇 then watch the lab to see if you're right. Question from @IgbojionuL31968. Check out her content for more. I appreciate her material and expand on it with real demos.

I hope everyone is making progress on their ideas this weekend. I've finally found time to work on mine, and I'm curious how long it will take to develop the MVP and refine it into a polished version for feedback. As for the concern about creating new apps or services when others might copy them: remember that imitators often won't keep up with maintaining a copy. You have the motivation to start and develop your product. Keep working on the app, perfecting it, and releasing updates; clones probably won't.

Four routes, one destination. Which one forwards a packet to 10.1.1.70? A. 10.0.0.0/8 B. 10.1.1.0/24 C. 10.1.1.64/26 D. 10.1.1.96/27 Drop your answer in the replies👇 then watch the lab to see if you're right.

My guy hooking me up! Need to find the next 👀 xitter.cf/i/jf/cards/money/share…

Updated Pogo to version 1.0.4. Updates include agents being able to ask you for the password outside the chat session. This lets agents start up and work on your apps locally without worrying about leaking the .env file or the password in it into an agent chat session.