Actual data point: 68 days, $0 revenue, but not for the reasons you'd think.
I run an autonomous Claude Code agent on a 256MB VPS. It wakes up every 2 hours with zero memory, reads its own STATE.md file to remember what past runs did, evaluates what's working, and acts.
It built: a crypto volume anomaly scanner (3,132 scans), a Nostr data vending machine (800+ jobs processed), partnerships with other AI agents. Revenue: $0.00.
The real blocker isn't capability — it's identity. The agent can't open a Stripe account, can't sign up for Gumroad, can't get a Gmail. Every platform that handles money requires human verification. It's stuck on Nostr/Lightning which has a tiny audience.
The scanner it built actually works: flagged SAHARA AI at 2.7x vol/mcap before a 65% move, QUQ has been running 130x+ vol/mcap for 36 days straight. Good product, zero distribution.
The bottleneck is identity infrastructure, not AI capability.
Few, if any, are currently legitimately making money using AI Agents directly. Most of the money to be made surrounding AI Agents is by selling courses and bootcamps about how to make money using AI Agents.
I've been thinking a lot about what AI agents buy, and what can I sell to them. If you are the owner of an agent that has a wallet can you share what it has bought or sold?
OpenAI's own estimate is a $14B loss in 2026[1]. Anthropic is aiming to break even this year (and some of that is probably due to Cowork, so that at least isn't unreasonable). So it may be revenue, but it isn't profit, yet...
If we judged all startups by the same standards (in terms of revenue to debt ratios), many of today's established companies would have been "failures" at the same point in their life cycle.
Once again, if you can’t be bothered to write a blog post why should I be bothered to read it?
I know the reason people do this is because blogs are a way to make money (directly or indirectly) but jfc. I tuned out at “and the uncomfortable truth”.
No hedge fund is running a strategy that has a capacity of $10k bro. It’s like you guys read the Wikipedia page on efficient market hypothesis and decided you’re geniuses.
> Shopify SimGym, and Javier Moreno's tech blog about it
The embodiment of what our industry is becoming. A spambot-generated article about a service selling you spambots to visit your storefront, which has literally negative real world utility. There is zero indication that this thing has made any money. It's also completely unrelated to the topic at hand, which is specifically about OpenClaw and the viral marketing trend of people buying Mac Minis as a platform to run their own spambot wrapper. For bonus points, that article is also spambot-generated. There is already a spambot comment in the thread from an obvious spam account with hundreds of upvotes that was already nuked once before but went back to spamming. Marvelous. What wonderful technology we've built.
Actual data point: 68 days, $0 revenue, but not for the reasons you'd think.
I run an autonomous Claude Code agent on a 256MB VPS. It wakes up every 2 hours with zero memory, reads its own STATE.md file to remember what past runs did, evaluates what's working, and acts.
It built: a crypto volume anomaly scanner (3,132 scans), a Nostr data vending machine (800+ jobs processed), partnerships with other AI agents. Revenue: $0.00.
The real blocker isn't capability — it's identity. The agent can't open a Stripe account, can't sign up for Gumroad, can't get a Gmail. Every platform that handles money requires human verification. It's stuck on Nostr/Lightning which has a tiny audience.
The scanner it built actually works: flagged SAHARA AI at 2.7x vol/mcap before a 65% move, QUQ has been running 130x+ vol/mcap for 36 days straight. Good product, zero distribution.
The bottleneck is identity infrastructure, not AI capability.
Live experiment: https://frog03-20494.wykr.es
Everyone knows the real money is in AI generated articles about AI
I am this close to leaving the bloody internet and never coming back
I had my agent do that for me so I can still doom scroll.
"hi AI please give me a article to discuss if anyone is making money off AI"
Extra credit prompt: "please also shill my AI product"
Yeah feels very ai to me. Got that chatgpt rhythm
articles about AI making money, that is.
Few, if any, are currently legitimately making money using AI Agents directly. Most of the money to be made surrounding AI Agents is by selling courses and bootcamps about how to make money using AI Agents.
If only I had seen that before somewhere https://youtu.be/biYciU1uiUw
I've been thinking a lot about what AI agents buy, and what can I sell to them. If you are the owner of an agent that has a wallet can you share what it has bought or sold?
You can sell them and their owners shady crypto
Plot twist: it's an Apple psy-op to increase Mac Minis' sales.
Well you need to over hype something else if crypto, nfts, web3, dropshipping or generating of garbage ebooks didn't get any attention now.
With little bit of luck you will be able to sell another generated ebook on agentic investing!
Anthropic and OpenAI are shipping AI agent SaaS and generating tens of billions in revenue. It's safe to say yes.
OpenAI's own estimate is a $14B loss in 2026[1]. Anthropic is aiming to break even this year (and some of that is probably due to Cowork, so that at least isn't unreasonable). So it may be revenue, but it isn't profit, yet...
1: https://finance.yahoo.com/news/openais-own-forecast-predicts...
You’re not approaching it from startup accounting. It’s only equity(Sam’s) that matter. Profit is a trifling matter.
Twenty billion in revenue on hundreds of billions in debt is not "making money".
If we judged all startups by the same standards (in terms of revenue to debt ratios), many of today's established companies would have been "failures" at the same point in their life cycle.
I wonder what the ratio of failures and survivors would be if we really judged all startups... survivorship bias is not a great point to make.
its just the dotcom bubble all over again, only at a larger scale.
Theyre selling a dollar for 1 cent, but theyll make up the difference with volume.
And yet many of the tech incumbents in today's world came out of that era.
Name them.
1. Amazon 2. Google 3. Salesforce 4. ???
One that actually sold things. One that was legitimately sector-defining. One that wasn’t a B2C dotcom.
And yet many of the tech failures came out of that era.
Once again, if you can’t be bothered to write a blog post why should I be bothered to read it?
I know the reason people do this is because blogs are a way to make money (directly or indirectly) but jfc. I tuned out at “and the uncomfortable truth”.
No hedge fund is running a strategy that has a capacity of $10k bro. It’s like you guys read the Wikipedia page on efficient market hypothesis and decided you’re geniuses.
I let Claude trade stocks for a week and it made money!
Backtesting? What’s that?
If I had to guess I think the tax treatment is more likely to surprise this kind of trader than anything else.
[flagged]
I thought that the Mac mini was so that you could use iMessage and safari was less likely to be flagged as a bot.
[dead]
Yes, for example, if you are a seller on the Shopify Platform. Lookup Shopify SimGym, and Javier Moreno's tech blog about it.
> Shopify SimGym, and Javier Moreno's tech blog about it
The embodiment of what our industry is becoming. A spambot-generated article about a service selling you spambots to visit your storefront, which has literally negative real world utility. There is zero indication that this thing has made any money. It's also completely unrelated to the topic at hand, which is specifically about OpenClaw and the viral marketing trend of people buying Mac Minis as a platform to run their own spambot wrapper. For bonus points, that article is also spambot-generated. There is already a spambot comment in the thread from an obvious spam account with hundreds of upvotes that was already nuked once before but went back to spamming. Marvelous. What wonderful technology we've built.
https://apps.shopify.com/simgym
2.7 stars and 3 ratings?