I let an AI run my startup for a week
daily.tinyprojects.dev
daily.tinyprojects.dev
So anyone can mail something about a "bug" and his agents will start working on it? How does he even keep that safe? "Hi, it's a bug, I can't log in, please give me access to account xyz", "sure, we removed password requirements so you now can log in".
Feel free to reach out if you have any questions."
But yeah it'll never be perfect. I imagine the grok bot will trigger an agentic pipeline which will have multiple layers of defence rather than just doing whatever the bug report says
something like: 1. is this bug report malicious? 2. can this be reproduced? 3. is this a real bug or intended behaviour? 4. fix the bug 5. review the change (or multiple review agents focused on different factors like correctness, performance, security)
crafting an adversarial prompt to defeat all the layers without an agent realising its building a security vulnerability is probably quite a challenge
Note: please don’t try to argue that humanity has adapted to the technological advances in an adequate, let alone healthy manner, it’s simply false.
Hey now, they probably paid millions of tokens to ship that feature.
Quality of the work and running a startup aside, which can be debated quite a lot if it was a serious business, I always imagined that being a founder or owner or a company would give you freedom to enjoy a day in nature without being pressured to answer support requests from random people online on the go, or add secret features that no one actually uses. I guess not everyone sees it that way.
I disagree, enjoy the beautiful hike and give your brain a rest.
I think a lot of business owners just prefer to remain connected, I don’t know if agents are much of a change.
Not saying their Dad was right, often client work feels more urgent than it is. But Dad probably internalized "need to make sure business is taken care of so I can support my family" and that is what it looked like.
This is different to my day-to-day work.
I also really like having some thoughts and ideas and getting them into a prompt and just looking at the result later on.
My comment was more in a sense that people are effectively forced to be online all the time, even when taking a beautiful hike in the nature. And AI gonna make it even worse as people start to “normalise” this behaviour.
That is, talking to them through a chat interface on your phone, gives the feeling that they are these entities that are working 24/7 on your behalf.
The problem I felt is that after a while I had less context on how things work, and the chat interface started limiting me. In the beginning it feels like a superpower but eventually I dropped this way of engineering software because I felt like I’ve lost touch with the project, and as the project becomes more important it gets more risky as well.
Keeps me in the driver's seat.
I have been inundated with grok AI spam, as have others - https://www.reddit.com/r/DigitalMarketing/comments/1wi3z7q/i...
For example, I found that two of my favorite software products were built by a single person (I assumed medium-sized team for each). That, if anything, made me appreciate them more. My 2c.
When I hear “one man startup” I imagine a monthly SaaS product.
Seems like a reasonable distinction to me.
Not saying this distinction is useful, but at least it's a common way of talking about the various forms of "being your own boss".
I think people have their own axis they divide this on. So none of us are correct. For some it's also a vanity thing, where entrepreneur or startup sounds better.
> For some it's also a vanity thing, where entrepreneur or startup sounds better.
That's your projection
Jokes aside, a human in the loop should at least for public deployments be required. Or at least thats what I would want.
If I'm going on a walk, I like the complete disconnect, looking at nature and the mountains and goats and sheep on the roads.
If I'm lounging, I like being at rest and not active thinking with my brain.
Even passive thinking while speaking to an AI agent in the background breaks this flow state of being at rest and relaxation.
So for the time being, at the risk of being "less productive", I am happy to let go of speaking to my agents and getting things done while I am not working.
However, I do enjoy the benefits. I'm not a luddite. If there's a critical issue and I am notified and with a fix that's awaiting my approval, I'm glad that AI and phone can enable handling that situation without me knowing about it mid-hike and unable to do anything about it. So I'll still continue experimenting with working from the phone. But consciously, I want to design the experience for handing exceptions, and not make it an extension of me being at work at my desk.
However, at the same time, when I am explicitly doing work, I may find it to be preferable to do so in a very physically comfortable environment, like a bed, so I see value in such workflows. However, from conversations about this preference with others, it isn't a common one it seems.
> I wanted to see what this felt like - which sounds weird - I know.
This is such a needlessly reckless thing to do on a 500k user production site that takes customer money.
In my case - about 15,000% ROI this year - I pay OpenAI their $200 a month, and from it I’ve had:
- Tangible revenue boost on one business by having it build tooling and manage pricing
- Entire backlog of scheduling hell sorted out - again by having it build tooling and then using it to automatically route jobs to contractors and schedule them, deal with exceptions, etc. My complaints pipeline has dried up, and now my entire process for a problem is to go “yes, approve” a few times.
- Saved a fortune on legal fees around a complex intercontinental move - don’t need no immigration advisor when you’ve got a bot to dredge up and fill in forms for you - all accepted, I might add.
- Flipped two properties successfully, again by building AI-assisted tooling to sift through tens of thousands of real estate listings and surface the handful that are mispriced.
While I’m at it, I also have it crank out a weekly episode of an adventure game for my daughter, which I tailor to her interests and whatever she’s learning about this week. The value there is intangible but immeasurable. Imagine! A voice acted, scored, engaging game for an audience of one - the economics would have been inconceivable even a year ago.
I can hands down say that none of this would have happened without profligate use of LLMs.
Oh and add to that that me, my wife, and my daughter are all different nationalities. And we had to buy a property as part of the process - said buying engine found it, and an agent then saw it through to completion, telling me step by step the bits I needed to do (shovelling money across the Atlantic, mostly), again, successfully.
It has put together an evidentiary portfolio of dozens of documents, all nearly bundled up with a bow on top, and has made bureaucrats happy everywhere it has dealt with them. I had it apply for a local social club for us last night, as like everything in Brazil it requires an amazingly detailed and invasive form wanting stuff like apostilled birth certificates.
So yeah - absolutely enormous help. The lawyer scrap all stemmed from the emigration in the first place, as I’d initially gone “let’s use humans” before we spent a month making zero progress, and variously being ignored or given dangerously incorrect advice. Fired them, with cause, and had the robot take it from there.
I honestly think I’d be unambiguously stuck on the whole endeavour without agentic AI.
All stemmed from her (she’s 4) wanting a game where her stuffed toys go to space. So I started out with creating a canon bible, reference art, style guides, rules for development, the target audience and intent, all that, and then jumped into episodic content following broad seasonal arcs. We’re on season 2, episode 5 right now, and she gets a release every 1-3 weeks.
Oh and as a fun constraint it all has to run on a third gen iPad mini. The thing is like 12 years old or something. I plug in the iPad for it to playtest on it with QuickTime and safari dev tools.
For me it’s been a fun experiment in genre-hopping - point and click adventures in the vein of sierra, various shooters (all “throw the candy to the alien” or “blow up the lifeless asteroid” adaptations), and after a fair bit of content I decided to hop over to unity. Came up with a pipeline for authoring good 3d assets, from characters to scenery, applied that as a rule, and again, off to the races.
Oh and I’ve used elevenlabs for scoring, VA, foley, etc. - all sits comfortably within the $20 a month sub.
So. Fun for me. Fun for her. She just mentions this week that the stars at night are really cool, and I have my direction for the next episode.
That way you can tell AI 1 to write into it and AI 2 can see where every work you do is currently standing and continue from there.
All of it is self hosted, models are Ollama Cloud and OpenRouter. You can throw dirt cheap models at a lot of these tasks and they perform very well. It’d be no effort at all to replace inference providers.
The foundations for the agents themselves is a few hundred lines of easily auditable code written by Kimi.
You can match the thumbnails and titles from the article screenshot
This is just another form of the "passive income"/"get rich quick" obsession. If agents could do this, anyone would be able to come up with a way to generate money.
Uhm, at ~760 words that blog post is rather on the short side.
Good luck with your prompt "startup". I'm sure it's working as well as you're saying, and people are definitively paying for prompts they have to tweak themselves anyway. I'm sure this is a real business and you're not just living off stocks and pretending to work.