Altman's eye-scanning startup is laying off. Here's the lesson
Tools for Humanity is cutting staff over weak revenue even as OpenAI files for IPO. What a billion-dollar hardware bet teaches small builders about money.
Tools for Humanity layoffs are a strange headline to read on the same day OpenAI files for an IPO. Same founder, Sam Altman, two very different stories. According to TechCrunch, the eye-scanning identity company is reportedly struggling to generate revenue and is cutting staff.
I don't run a hardware startup. Neither do most people reading this from Sri Lanka. But the gap between "famous founder, big vision" and "can't make money yet" is the most useful thing in this story, so let me pull on that thread.
🔍 What's actually going on
Here's the short version, with only what the source supports:
- Tools for Humanity is Sam Altman's identity-verification company (the one behind the iris-scanning orb).
- It is reportedly struggling to generate revenue.
- It will downsize its staff.
- This is happening as OpenAI files for an IPO.
Key takeaway: A founder can be at the center of the most-hyped company of the decade and still own a sister venture that can't pay for itself. Hype and revenue are different machines.
I'm not going to guess headcount numbers or funding figures the article doesn't give. What matters is the pattern, not the spreadsheet.
💰 Vision is cheap, revenue is the hard part
The orb is a genuinely ambitious idea: prove a human is a unique human, online, without a password. The problem is that "ambitious" and "people will pay for this today" rarely arrive at the same time.
Compare the two bets under one roof:
| Venture | What it sells | Revenue signal |
|---|---|---|
| OpenAI | API + ChatGPT subscriptions | Strong enough to file for IPO |
| Tools for Humanity | Identity / proof-of-personhood via hardware | Reportedly weak, now cutting staff |
The difference isn't talent or funding. It's that one product had a thing people swipe a card for every month, and the other is still convincing the world it has a problem worth paying to solve.
If your idea needs the world to change its behaviour before you earn your first rupee, you are running a research project, not a business. That's fine, as long as you fund it like one.
🛠️ What a solo builder in Sri Lanka should copy instead
I build small tools. The whole point is that they earn or save something fast, on a tiny budget. Here's the playbook this news reinforces:
- Charge (or get used) before you scale. One paying user beats a thousand on a waitlist.
- Avoid hardware until software proves demand. Hardware burns cash on a fixed clock. Code doesn't.
- Keep your burn low enough to be patient. A free Vercel tier and a $5 VPS buy you years; a factory buys you months.
- Solve a problem someone already googles. Don't teach a new behaviour and sell a product at the same time.
For the money side specifically, if you freelance or sell anything in USD while your costs are in rupees, run the numbers honestly with our freelancer USD to LKR calculator before you commit to a runway. Knowing your real take-home is the cheapest risk control there is.
🌐 The "proof of personhood" problem isn't going away
Here's the twist: the problem Tools for Humanity is chasing is real and getting worse. As AI writes more comments, reviews, and accounts, "is this an actual human?" becomes a serious question for anyone running a site.
But the lesson for builders is about scope, not surrender:
| Approach | Cost | Realistic for a small team? |
|---|---|---|
| Scan everyone's iris with custom hardware | Enormous | No |
| Email + phone verification | Low | Yes |
| Bot-check on signup (Turnstile / hCaptcha) | Near zero | Yes |
| Rate limits + behaviour heuristics | Free, just code | Yes |
You don't need an orb to keep bots out of a comment box. You need a free anti-bot widget and some sensible limits. The expensive moonshot version of a problem is rarely the version you have to build.
I use exactly this stack on my own site. A bot-check on signup, rate limits on the API routes, and email verification cover almost everything a comment-and-tools site needs. The cost is a few hours of code and zero rupees a month. The orb's job is to solve the same trust problem at planetary scale with no shared infrastructure, which is a genuinely harder thing and explains why it costs so much more.
Bottom line: Pick the cheapest tool that solves 90% of the problem. The last 10% is where startups go to die.
📊 Layoffs are a signal, not a verdict
A round of layoffs doesn't mean an idea is dead. It usually means the spending got ahead of the income. That's a correction, and small teams have an advantage here: we never had the headcount to over-commit in the first place.
Things I read from this, and try to apply:
- Match your team size to your revenue, not your ambition. Ambition is free; salaries aren't.
- A famous backer is not a business model. Distribution and a price tag are.
- Optionality is wealth. Staying small keeps the option to pivot. Staying lean is what lets you survive the year your big bet doesn't land.
💡 What this means for you
If you're a student, an engineer, or a two-person team in Sri Lanka watching this from the outside, the takeaway isn't "Altman failed" because OpenAI clearly hasn't. It's that even unlimited access to capital and talent can't fast-forward a market that isn't ready to pay. You and I have far less money, which forces the discipline these big bets sometimes skip: earn early, spend slow, build the cheap version first.
So when your next idea feels like it needs special hardware, a new user habit, and a year of runway before its first sale, pause. Ask whether there's a boring version that charges money this month. Build that one. You can always grow into the orb later, once something is actually paying the bills.