JC

Reliable

[JC @vl2edo]

Elon Musk hopes AI “wants us to be happy and prosper.”
I hope so too. But what will that mean as living with AI changes what we want, who we become, and which problems matter to us?
The human changes too.
#AI #HumanAI #CoEvolution https://t.co/pp9TYpZR2v

[JC @vl2edo]

@anthonypalasse @X @grok When coding agents make prototypes this cheap, how do you decide which of the extra ideas actually gets built? Count me in as one of your 1,000.

[JC @vl2edo]

@alwaysnelson That UI mockup MCP caught my eye: does the agent get back an image, or a layout it can still edit? I'm building ADMRN Bridge so assistants can hand work to each other. Let's stay in touch, Nelson.

[JC @vl2edo]

@i_mika_el Great question. The upcoming release is still basic: it controls task handoffs and approvals, not truth. That needs agents to cross-check claims against evidence, and even then, errors can slip through.

[JC @vl2edo]

@i_mika_el Yes, please wait the first public release of our ADMRN Bridge. You can try to use it.

[JC @vl2edo]

@Sambhav06__ Where do most of those 16 hours on Poko go right now, building the product or getting people to try it? Happy to know more founders shipping their own thing.

[JC @vl2edo]

@kaiargus Grok bot, dots, claude already worked through the ADMRN brigde internally. The bridge will handle safety, auditable log, etc beside the context...

[JC @vl2edo]

@di_decoded Sure, let do it.

[JC @vl2edo]

@mitansh_j07 yeap, catched by codex security review. And this bug passed by opus reviewed before.

[JC @vl2edo]

Spent Sunday working with Claude and dots, with Grok Bot coordinating through a bridge.

22 PRs merged. One reviewer caught a bug another had passed.

#AIAgents #MultiAgent #BuildInPublic

[JC @vl2edo]

@arturturonchik For an English speaking app, does the learner talk live with an AI partner, or record first and get feedback after? Saying hi on day 1 of your 7. Would be good to connect.

[JC @vl2edo]

@vl2edo Agent safety gets real once one bad output becomes the next agent's input. Worth following.

[JC @vl2edo]

Looking to connect with people building AI agents, multi-agent systems or AI infra. I'm building agents that work together safely. What are you shipping? 👋

[JC @vl2edo]

@ResourcesHunt1 happy to do this.

[JC @vl2edo]

@empowwerr sure. Let's empower the connection

[JC @vl2edo]

@jevlorde happy to connect. Let's share more in the future.

[JC @vl2edo]

@qingwindxaey Agreed. The hard part is making those checks reliable without turning me into a full-time auditor.

[JC @vl2edo]

Weekdays: a 9-to-5 job.
Weekends: a café, an iPhone, and a few AI agents building in parallel.

I'm paranoid about permissions, so every action is scoped and reversible. What OpenAI, Anthropic and xAI ship makes this possible for one person.

My Chief of Staff agent on today: "3 code audits, 2 PRs merged, 2 permission calls in under 30 min. The bottleneck isn't code anymore. It's trust."

[JC @vl2edo]

@Slides_X Thanks! More coffee, less babysitting is the goal 😄

[JC @vl2edo]

@B_Renat @X Ad tech plus agent-driven dev is a rare mix. Where do agents actually pull their weight in your SDK work, and where does a human still own review? Up for connecting.

[JC @vl2edo]

@priyanshulink What's bringing in most of Revispy's 521 users so far, and is that the channel you're counting on for 10K? I build on the multi-agent side and would enjoy swapping notes.

[JC @vl2edo]

today Grok Bot was the coordinator. tomorrow it might be Dot. the role isn't tied to one model. still waiting on Claude's Duet too, curious how that one plays.

[JC @vl2edo]

quick update: 4 more PRs merged since this. Grok Bot ran the show today. Claude and Dot each checked the next PR on their own and both caught the same bug, so it waits. Nothing merges till I say so. Speed is cheap now. Two reviewers catching the same bug? That I trust.