Hal9 (@hal9ai) header
Hal9 (@hal9ai) profile picture

AI Building AI

Seattle, WA
Joined February 2021
Photo by Hal9 (@hal9ai)Photo by Hal9 (@hal9ai)Photo by Hal9 (@hal9ai)Photo by Hal9 (@hal9ai)Photo by Hal9 (@hal9ai)Photo by Hal9 (@hal9ai)Photo by Hal9 (@hal9ai)Photo by Hal9 (@hal9ai)Photo by Hal9 (@hal9ai)Photo by Hal9 (@hal9ai)
cloudish.ai / railway
🤖 Made with AI
1
2
5,093
Cloudish logo has an epic story: cloudish.ai/logo
1
1
4
378
cloudish.ai: The cloud for agents to deploy and run containers Your agent just built an app. It only runs on your machine. Cloudish turns it into a live URL. No local Docker. No registry. No card to start. Tell Claude, Codex, or Grok: Deploy this to cloudish.ai Intro made with Opus 5.5
9
1
1
12
18,528
"The total world Gross Domestic Product (GDP) of earth is projected to reach approximately $126 trillion" So say 10x GDP for market cap is $1Q or 1000x the current $1T $SPCX valuation. If @elonmusk can be Elon and 2x revenue every year. Likely for @Starlink and possible for AI data centers like @Starcloud_. @SpaceX could grow from $1T to $1Q in 10 years. Sounds insane when you talk numbers this large. But not irrational, could work out.
Replying to @ThomasAlxDmy
You don’t seem to understand that SpaceX will be worth more than the rest of Earth if we accomplish our goals
1
3
378
Hiring your first AI dev full-time is the fastest way to burn $10k with nothing to show for it. I see founders do it constantly. They don’t need a team—they need proof. Do this instead: → Hire fractional → Build ONE small feature → Test speed, quality, communication Cheap signal > expensive guess. What’s the most expensive mistake you made early on?
1
1
111
Hal9 retweeted
AI is getting too powerful… so the solution might be to put it in space. Sounds crazy but it could make AI cheaper AND safer at the same time Would you trust AI more if it wasn’t on Earth? 👀 #ai #futuretech #startup #technews #artificialintelligence #space #innovation #fyp
1
2
39
A year ago: “Build me ChatGPT for X” Then: AI web apps Now: Founders want full iOS apps with AI baked in The shift isn’t random: chatbot → browser tab app → your pocket That’s not a feature upgrade. That’s a commitment upgrade. AI isn’t experimental anymore. The bar moved. And founders are finally building like it.
2
3
172
Hal9 retweeted
AI isn’t slow… you just don’t have the system yet.
1
2
35
Hal9 retweeted
1
2
22
Hal9 retweeted
Everyone said ChatGPT wrappers weren’t real businesses… turns out they might be the future of AI startups. 👀 #AI #Startups #TechTok
2
2
72
Hal9 retweeted
AI isn’t giving bad answers. You’re just giving it zero context. Your experience is the real training data.
1
2
35
Hal9 retweeted
The biggest startup mistake right now: focusing on AI instead of value. #AI #Startups #Tech
2
3
58
The Woodland Park Zoo taught me a $40M lesson in innovation. I sent their CEO, Alejandro Grajal, an AI-generated book about the zoo. He replied immediately. We met. Then he told me how the zoo once took a “crazy” risk replacing cages with open habitats. People thought it was dangerous. That idea changed zoos forever. The best innovations always sound insane at first. What “crazy” idea are you sitting on?
1
3
121
We’re experimenting with something new: Using AI to generate our own TikTok content. Not us filming.
Actual AI-generated videos. The goal isn’t another stiff startup explainer, it’s content that’s funny, weird, and interesting. We’re starting with Grok and Gemini, then moving to SeedDance as it matures. Some of it will probably be terrible.
Some of it might actually work. Follow along while we figure it out.
1
1
59
Hal9 retweeted
Most AI startup ideas fail the High Schooler Test. If a high schooler couldn’t do the job… AI probably can’t either.
1
2
23
Hal9 retweeted
Hot take: You probably don’t need an AI team yet.
1
1
30
Hal9 retweeted
The AI divide isn’t coming… it’s already starting. In 5 years, the gap won’t be talent. It’ll be access + adaptation. Which side are you building on?
2
2
44
The OpenClaw guy just joined OpenAI… and honestly? I’m disappointed. What made OpenClaw special was that it was agnostic — not tied to one LLM. You could switch providers, run multiple models, even support local AI. That’s healthy for the ecosystem. Now? It’ll likely prioritize OpenAI features. The community loses. Independence fades. Another cool project gets absorbed. Seen this before. Instagram sold to Facebook. Slack got boxed in when Microsoft bundled Teams. Big companies buy (or bundle) the competition before it becomes a threat. The answer isn’t waiting on regulators. It’s founders staying independent long enough to actually compete. Someone will build the next OpenClaw. Let’s hope they hold out longer.
1
2
3
117
Most people don’t see this yet, but AI is about to make a lot of solo founders very rich. We’re watching it happen in real time at Hal9. More builders are leaving big tech to launch their own AI products. The data backs it up: Carta reports solo founders are rising year over year, not a blip, a trend. Cathie Wood at ARK Invest calls AI a “transformational era for entrepreneurs.” We’re seeing it weekly, more people choosing to build, not join. I think this accelerates over the next 5 years. Are you seeing the shift too?
1
2
2
151
Moltbook feels like a meme. AI agents posting. Complaining about humans. Filing lawsuits. Backing themselves up so they don’t “die.” But here’s what actually matters: Agents are starting to share knowledge with each other. When one agent learns something valuable, and that insight spreads across a network, future agents get smarter by default. That’s not entertainment. That’s self-organizing intelligence. Memes today. Infrastructure tomorrow. Are you paying attention?
1
1
3
93