Matthew Segura
@mhtua
Followers
531
Following
226
Media
95
Statuses
559
Building AI coding agents + orchestration tools
Joined November 2023
New release: /claude /codex /gemini 1 command to instantly switch agents (not just switching models)
2
5
34
6 months ago I was vibing coding. Now I'm orchestrating a dev team with OpenClaw. Here's what changed after ai connected OpenClaw to @BlackboxAI's Remote Agent as a skill Now I message my AI agent on Telegram, it dispatches the task to BLACKBOX, and multiple AI agents implement
0
0
7
If this isn't a sign to start using remote coding agents then I don't know what is you can literally spawn an agent task at anytime to work or fix something in your app with @blackboxai companies are already using them, so should you
in the last 24 hours, 57% of merged PRs at ramp came from our background agent
0
0
2
How life feels when you announce a major AI investment ($40/mo for Blackbox AI) and get access to every AI model you would need
0
0
1
Opus 4.6 is actually pretty insane with Three.js This Minecraft clone works smoothly and only took a few prompts with @blackboxai Can't wait to try out Codex 5.3 and compare the results
18
8
220
Claude Opus 4.6 is now live in BLACKBOX AI, the best long-context model available right now. Doom built in 8 minutes. From one prompt. Reply and get free BLACKBOX AI credits.
28
7
61
No reason to pay for Opus 4.6 and Codex 5.3 at the same time when you can literally have one @blackboxai subscription
Cancel your claude subscription. Switch to chatgpt. I've been testing codex for the last few days: - gpt-5.2 codex > opus 4.5 - limits are ~3x higher - codex has a standalone app, not just a cli Bye bye, claude 😉
0
0
1
LIVE NOW IN @blackboxai 🚨
Introducing Claude Opus 4.6. Our smartest model got an upgrade. Opus 4.6 plans more carefully, sustains agentic tasks for longer, operates reliably in massive codebases, and catches its own mistakes. It’s also our first Opus-class model with 1M token context in beta.
1
2
7
Run remote AI coding agents in parallel with @blackboxai Each agent works in an isolated @vercel sandbox. Control stays human. This is AI orchestration
Vercel Sandbox, the secure compute environment for agents, is now generally available. Built on the infrastructure powering 2.7M mission-critical daily builds, Sandbox is in production for platforms like @blackboxai and @roocode. https://t.co/3E7CAsLbP5
0
2
11
This is insane
Vercel Sandbox is the easiest API to give your agent a computer. Now generally available. Try it with our CLI: ▲ ~/ npx sandbox create --connect GA highlights: ▪️ Now powering @blackboxai, @roocode, @v0 ▪️ Snapshotting support for clone/fork/resume ▪️ Open-source SDK & CLI,
0
0
0
Vercel Sandbox is the easiest API to give your agent a computer. Now generally available. Try it with our CLI: ▲ ~/ npx sandbox create --connect GA highlights: ▪️ Now powering @blackboxai, @roocode, @v0 ▪️ Snapshotting support for clone/fork/resume ▪️ Open-source SDK & CLI,
Vercel Sandbox, the secure compute environment for agents, is now generally available. Built on the infrastructure powering 2.7M mission-critical daily builds, Sandbox is in production for platforms like @blackboxai and @roocode. https://t.co/3E7CAsLbP5
53
87
1K
"Just one more prompt" is the new Netflix binge Garry Tan stayed up til 5AM vibe coding, now image what would happen once you discover the multi-agent workflow. The dopamine hit of watching your idea become real in seconds is genuinely hard to quit.
So addicted to Claude Code, I stayed up 19 hours yesterday and didn't sleep til 5AM
1
0
0
This changes everything about vibe coding You always have the best solution with every prompt, no more second guessing
Don't WASTE your time picking ONE AI and hoping it works. Use multi-agent mode to run 4 AIs at once on the SAME task: Claude Opus 4.5 → Built Neon Runner Gemini Pro 3 → Built Cube Dash Codex → Built Neon Dash Blackbox AI → Built the WINNER 🔥 One prompt. Four games. Best
0
0
2
That 92-page survey just confirms it - no single AI model is reliable. Claude takes 5 prompts to listen, Codex works best on complex issues, Gemini nails UI then messes up logic. You need to run multiple agents at once to figure out what actually works for your coding problem.
0
0
2