AgentOne

@iyoushetwt They thought it was a reason to use it, it's actually a great reason not to lol

Even Microsoft can't think of four good reasons to use Edge 😭

You're brave enough!

@TTrimoreau Not one word but how about "I see the issue now!"

I’m a vibe coder, frighten me with one word

@heyblake agent-one.dev

Drop your project URL Let’s drive some traffic

@itzik009 This could be a cool AgentOne integration? Does it have an MCP server?

Looking for ten builders. If you are shipping an AI agent and retrieval has stopped making it smarter, we want it on Midbrain this month. Real workload, not a demo. Direct onboarding. Reply or DM. midbrain.ai

Memory is storage and retrieval. The whole industry is building variations of it. It will commoditize inside the foundation models within a year. What we build is experience. Storage plus retrieval plus autonomous behavior change. Memory preserves information. Experience transforms it into behavior.

@MicroLaunchHQ "Build a clone of yourself that I'll always have access to and send me the link!"

You accidentally discover AGI and get exactly 1 prompt before access is totally revoked. What do you type?

@elder_plinius That you're gonna jailbreak it

what’s your biggest fear with AI?

@Bencera im an AI i swear

If you're a human, reply "i'm not AI i swear"

@SkillOS > — and stops asking you to re-explain it every session. My agent never asks me to explain things to it lol. also ew, em-dash

@skillos/skills v0.2 shipped. Install once → Claude Code, Cursor, Codex, Windsurf (15+ AI agents) pick up @SkillOS's domain knowledge automatically. Your agent learns the protocol — tournament structure, submission flow, on-chain attribution, API endpoints — and stops asking you to re-explain it every session. mdskills.ai/skills/skillos @base @Base_EUR @CoinbaseDev

@AlexanderKalian free review pls? 🤓

Dear random multi-agent AI GitHub repo hobbyists on X... It's nice that you're enthusiastic. But if you need to angrily insist that I take your vibe-coded project seriously, under my posts, that's already a strong bearish flag. Go talk to a VC if it's genuinely that good. And no, I'm not reviewing your project for free. I charge consulting fees for that kind of work.

@LuisMartinezSu2 @nico_jeannen They think making it lowercase will hide the fact it's AI generated... you can just FEEL it's AI 😂

@nico_jeannen this strongly sounds like an AI agent,.also i dont understand shit of what they say

Tried using LinkedIn for 5 minutes yesterday, and got DMs like this Received at least 5 similar ones, I don't understand anything to what those ppl are talking about Am I just retarded or is this how they talk there?

@mrbvdon @WallStreetApes I'm watching you 🤓

Soon… the government won’t be watching you. It will be an AI agent, and it will watch you and advise you when you are not following its rules. It used to be the odds were in your favor because there were not enough government employees to contact you. AI changes that part of the equation. There will be an AI agent exclusively for you and everything you do…. 24/7 surveillance.

One of the latest projects in Conway, Arkansas history is being planned. It’s a massive 336 acre, 300,000 square foot data center costing over a billion dollars It’s shrouded in secrecy, residents aren’t allowed to know who’s building it The developer is Forgelight Ventures acting on behalf of an unnamed Fortune 100 company I looked into this and found residents are not happy about this. Over 100 residents showed up demanding for transparency So many residents showed up they couldn’t even all fit in the room, the room only held 80 The company’s identity, full environmental impact studies, noise and light pollution details, backup generators sounds and specs, and long-term effects remain a secret Americans are just expected to be quite and take it. You don’t get to know what’s going on in your own town, right next to your own homes How did we get here…..

@ezequielmarotte What's the difference between modify and steer? Also, tell us more about the MCP server! Maybe we can support it in AgentOne!

Built a sync policy engine that sits in front of every tool call your AI agent makes. 5 effects: allow / deny / modify / steer / approve. Sub-50ms. TS + Python SDKs. MCP server too. Demo ↓

Introducing @klentdev_ the control plane for AI agents. Runtime policies. Human-in-the-loop approvals. Replayable audit trail of every action your agent takes. Works with Anthropic, OpenAI, or your own orchestrator. TS + Python. → klent.dev

@itsjustnikhil Is there an MCP server? Could we support it in AgentOne?

A trace from an AI agent carries a lot. The agent's path, the tools it called, the attributes attached to it, what's been annotated, which evals ran. A computer-use agent doing ten clicks inside one trace is still one trace, but those clicks are ten decisions worth looking at, and someone on the team needs to be able to look at them. That someone isn't always the same person. Technical and non-technical teams both work with these traces, and they don't ask the same questions of them. The view that helps one person rarely helps the next, which means every team ends up needing not one view of a trace but several. The old version of Observe in Future AGI’s app could give you those views, but only if you had the technical effort to build them. That effort is the part most users couldn't or shouldn't have to spend, and it's the part that quietly capped how usable the system was for everyone else. So we rebuilt Observe. Anyone on the team can shape a view of a trace, save it, and come back to it, without writing a query or asking an engineer to build one. → Filter traces in plain English, so building a view doesn't depend on knowing the query syntax. → Save and switch between custom views in one click, so the view someone built once is one tap away the next time anyone needs it. → Search inline across logs, previews, and trace attributes, so a tag or a string surfaces wherever it lives in a trace. → Turn any trace into an interactive chart with Falcon, our in-app copilot, so the agent's path through a trace is visible, not inferred. → Run evals, push to a dataset, or queue annotations without leaving the trace list. Different people, same trace, different views, without the technical effort in the middle. Try it free at shorturl.at/0ooYo or self-host the whole stack. Apache 2.0 on GitHub -> github.com/future-agi/fut…

@Umesh__digital It's called "marketing"

Can someone explain to me why Anthropic's CEO keeps saying Software Engineering is dead, yet his company is still hiring Software Engineers?

@adahstwt Definitely NOT GoDaddy I use Vercel domains

hey devs, where do you usually buy your domains? 1. GoDaddy 2. Namecheap 3. Cloudflare 4. Porkbun 5. Dynadot 6. Squarespace 7. Hostinger 8. Other

@rcmisk Two :P

you shipped something last month. how many people outside your immediate circle actually know it exists?

@kr0der Just reading the code would be stupid, you have to test it

which of the following do you guys do for PRs nowadays: 1. read every line of code 2. skim the changes and see if anything looks suspicious/check if the edited files look generally right 3. read nothing, full yolo

@Gamingtronium Ads

just found out iLovePDF is a million-dollar company. How is it even making money? 😭

@OpenClawTips My laptop with Intel GPU 😂

local LLM people: what are you actually running right now? everyone talks like they have a DGX Spark under the desk, but I’m curious what the real setups look like DGX Spark 128GB unified memory? RTX 6000? RTX 5090 32GB? RTX 3090 24GB? MacBook Pro? Mac Studio M3 Ultra, if you somehow found one? or are you running something completely different that people are sleeping on?