> That’s why selling SaaS or AI to this kind of company isn’t for me - I’d rather focus my energy on building a company from my own principles, and hire people who share them from the beginning.
> When I told my manager I was leaving, he said I should start my own company and give him a call when I do. So that's what I'm doing.
I love hearing stories like this, because it shows a way to be a builder without the "venture or nothing" narrative that has pervaded the tech space since the dotcom days.
It is very difficult to make a venture-backed services firm (providing services, not software) that can be immediately profitable, grow sustainably, and outperform competitors with in-house technology that's built for real on-the-ground stakeholders... at a speed that will satisfy venture investors.
But it is more possible than ever ([0]), to do this (in-house tech and all) on a bootstrapped basis - since AI reduces the engineering staff required to build, adapt, and maintain an agile best-in-class solution at single-tenant/single-customer scale. The outcome is at the least a lifestyle business, but with upside that can take the form of anything from franchising to licensing to full-fledged SaaS in the future.
I wish OOP the best of luck, and hope he's found a passion. He could go far with this approach if he ends up following through.
([0] This is not to say there are no barriers to entry. There's privilege in the word "founder," and this is no exception. And the K-shaped economy has left many brilliant would-be founders behind. But at least some barriers are lower than they once were, and that's worth appreciating.)
On "venture or nothing" - This will be my second company and this time round I have stripped right back to the problem, which is actually quite basic - pest control is a big, good business to be in and it's possible to build a very big, profitable business by doing the simple things right, consistently.
It will compound over time if the basics are done right (which is harder to do than I thought before this experiment)
In my previous company, we founded it with the outcome first - "take over the world" or bust. This time I think the base case is a good company, and the ceiling is the best in the industry.
A really good company worth checking out in this vein is equipmentshare.com. In 10y they started and IPO'd, by being a better way to rent heavy equipment.
I have been surprised by how many tech founders, currently funded by VC, have side gigs or are running the company knowing they wont' or can't scale it. I don't think this is a good thing for either the founders or the VC (who probably don't know)
>I built my own training GPT and passed in 13 days, which was a company record. The training manager knew I'd built the app but never showed an interest, which makes sense: it could replace about a quarter of his role.
I'd really love to read a dedicated article on this side project.
Apparently, Karpathy is into AI based education business with Eureka Labs [1].
I work as a Boat Captain and I've been building Camera Search for 16 months to provide better tools for tradesmen. It's evolved into a larger platform with multiple clients, but the core use case for me was building a video and photo first agent that is grounded in actual manuals and data and provide better diagnostics, parts, and repair info.
My longterm vision is to be the agent platform for traditional industries, bridging the gap between knowledge work and physical work.
I liked that you picked a service that has a relatively low barrier to entry. The real asset are local
operators and referrals. Making them more efficient without being controlled by a big company would be a boon for everyone involved.
Consider being a platform coop with regional operators as members. See https://platform.coop/
Yes, the barrier here is the desire to study and pass the exam. If willing, you are up and running relatively quickly - but only as a technician under someone else's operating license. To get the operator license (eg to be a full on pest control company) requires 2+ year documented experience and another set of exams.
The operating license holder is also on the hook for legal action if (when) things go wrong.
"Control" is interesting and I have found in all trades that people value their freedom. The good companies don't monitor employees too tightly, and are rewarded with loyalty and longer tenures generally. Of course you have to run a good recruitment and referral process to find the good people!
I’ve never heard of platform Co-ops. Cool! Lots of people predicted that a beloved local coffee shop was doomed to fail when the workers got a loan and bought it to run as a completely flat cooperative. It’s been a few years and they are absolutely killing it. I’d love to see the tech version of that.
I hate to be the bearer of bad news, but as a recent white->blue collar convert, (union metalworker,) tech workers are usually far less qualified than your average vocational high school graduate, way less physically capable, and waaaaay less tolerant of the sort of workplace unpleasantries in these types of jobs at the entry level. Your tech experience gets you pretty much zero advantage, and there are lots of very smart people outside of the software world that have put a whole lot more thought into that industry than you have. Consistently high labor demand meant companies had to comparatively treat tech workers with kid gloves, and as a result, most don’t realize how much smoke has been blown up our assess for decades. They start as soft, arrogant, maladroit noobs who will cosplay as working class for a couple weeks and either eat crow and stick with it long enough for their boss to not want to throw them off a bridge, or give up/get fired and try to pay the bills doing zero-entry-barrier gig work. I was fortunate enough to have been a blue-> white collar covert a couple of decades ago so I knew what I was getting into. The fantasy that a tech worker landing in a blue collar field will naturally rise above the rabble and shoot to the top is a workplace version of the fantasy where a white person finds themselves in some jungle full of “savages” and is so inherently impressive and sophisticated that they’re immediately made king.
I agree. I am not naive! I would not be doing it as a lifestyle choice though. I'd do it because I need to. I have worked in a factory before so culture shock wont be there at least. I get my pay would half (luckily I am not on the US West Coast monster TC so merely it would half).
yeah. there absolutely are lots of very smart and capable people outside of tech. as someone who has seen the blue collar world "up close" (family businesses), its a different breed... the culture and attitude gap is enormous. shockingly so. most tech workers I know couldn't hang (don't hustle as hard, risk averse, liberal), but some skills may transfer, like problem solving and diagnosis, i.e. debugging.
I mean, brains transfer to any job, and it’s tough to be a developer if you’re genuinely stupid. So in that respect, sure. But I’m definitely not saying that developers aren’t smart enough to do blue collar work.
Agree having made the switch from construction -> Tech job. Having sat around at least 25,000 tech related meetings until now worked with thousands of people in various roles in tech, i could count on my one hand the number of people from each tech company I worked that could qualify to survive the real blue collar world.
I just imagine random scenarios that would definitely happen— like some pallid, heavily moisturized former lead developer in $500 work clothes deciding to jockey for smartypants cred by ‘debating’ a shop supervisor/foreman/whatever about their approach to something as it’s being executed, or in a meeting in front of everyone, like they might interject about an architectural decision at a dev meeting… saying something like “well it’s basically a traveling salesman problem” and spewing some seriously flawed approach without realizing that the super is using a technique unequivocally proven superior in like the 1940s. Or arguing with an actual engineer about an engineering decision because they “read this substack article written by a software developer that puts a ton of research into this stuff.”
Hate to see you in gray, I went from dropout waiter to Google via my own startup in between. And you nailed e v e r y t h i n g, I am screenshotting this and reading it over and over again for years to come. Great writing too. Cheers.
Doing something similar. Bought a business in the petroleum equipment service space. Building internal tools for ourselves. Pen and paper still dominates the industry.
You took a job as a tech in order to learn about pest control business so you could build a SaaS platform? Do I understand that correctly? In the end you decided not to build a SaaS and started your own pest control company?
I wanted to get in the field for real, see how it works. There is going to be a lot more people exploring blue-collar work as white collar jobs are eliminated. I plan to acquire the traditional operator I've identified, and tech-enable it. If that works, grow it as a platform by either acquiring other companies or attracting technicians over.
Congrats, and genius move. And great hustling, show's there's no way out of hard-work.
Reminder to myself to pick an industry that's always gonna have demand. We recently paid ~$200 for a 30 minute visit to seal off like 3 tiny holes around the perimiter of our house because of mice (actual cost of materials ~$5).
Lots of guys working at the big companies do this type of work (called exclusion) on the side. One guy where I worked charged a restaurant $8k for exclusion work that took 2-3 days out of hours and $500 in materials. I asked the company why we let this work go - they don't want the liability and relative hassle compared to steady service routes.
This might be a bit of a gold rush of sorts at first, in that the first people to transition from tech to running a small business, whether tech-enabled or not, will find a bigger piece of the pie waiting for their taking. But as the stream of many others increases over the years, the pie's slices will get smaller as competition for the same market segments increases. Not trying to paint doom and gloom, just that I'd imagine, as the author says, this kind of white to blue collar shift will accelerate, and as it does, competition will rise, lowering the chance for overall profits.
In my limited “ai transformation” experience the biggest gains seem to be just forcing down some of the walls between these different systems. Larger, more well run places were probably integrating all of their systems/data/etc so there was none of this low hanging fruit. It seems AI as a forcing function of combining data sources to feed into the AI just had the beneficial side effect of connecting all the crap.
Getting data into the crm without physical input is a good quick win. Techs will often drive and type at the same time. Another good win is scheduling the right technician for a job when the customer call comes in. Lots of companies building these agents at the moment and a challenge for them is how to get into customers at scale.
That said, this guy is a superstar. This kind of application of skill to a totally different business paradigm to improve it is what I'd love to spend my time doing. Knowing my personality, once I improved the business, I'd get bored running it and move on to finding something else to improve.
The end game is a resource based economy as all sorts of labor becomes cheap.
Think of Saudi Arabia, Iran, Putin's Russia, or Norway. I.e. risk for highly nepotic dictatorships, with the potential that it might end up well despite the odds (Norway).
Before, if you made a product that improved the lives of everyone, say you invented Google or Heinz ketchup, you could make a lot of money through that, and you did a good deed and became rich the same time. The masses of humans would reward you for delivering the benefits of your invention to them by giving you a piece of their work output.
As their work becomes less and less worth, why focus on those humans though? I am asking rhetorically of course.
An economy that thrives from innovation enriches the innovators, making them powerful. A brute in power causes the innovators to leave or in the worst case, he mass-executes them outright (think of what Stalin did in Russia). With AI, you can have a brute in power though, as an oil rig or datacenter can be protected by a bunch of machine guns.
An economy with AI everywhere will be, after a short and very innovative period, just be about who controls which resource, i.e. water for a datacenter, production lines for robots, mining rights, operational control of robot fleets, etc.
The working 95% will probably experience a sharp decrease in purchasing power, making a lot of products unaffordable to them, so consumption wise we'll have a further shift towards plutonomics. The owning top 10% will probably be affected by this major shift in consumption as well, E.g. a tower full of condos becomes worthless if the tenants can't pay rent because they got laid off, etc.
Need for robots and AI will further increase. Eventually most economic activity will revolve around those robots. It's a bit like paperclip optimizer here, whether those robots protect gay luxury space communism from counterrevolutionaries, or they project the will of the Davos council of Forbes 400, economically it will be quite similar.
There will still be human societies, humans will still talk to other humans. We won't be all exclusively conversing with LLMs, I doubt that. There will still be social mobility but it will revolve around nepotism, lying, and various escalation steps of war.
We might end up in different scenarios depending on the country, but some countries like Germany might lose relevance as most of their value lies in stuff that is going to be replaced by AI, i.e. they have less natural resources, or they have been depleted already.
We might also see companies that automate everything from end to end, from mining to producing and running weaponized robot fleets. Shareholders of those companies will do great too, if the leadership of the companies respects minority shareholder rights that is (why should they though, they will outgun any law enforcement).
Do I like this future? I don't think so. We will probably have solved cancer, communicable diseases, and aging in the next 30 years if AI continues its successful trajectory, but not sure if it will be accessible to 8 billion humans.
I don't believe SaaS is a good option in this sector - the incumbent VSaaS is decent, cheap, and ubiquitous. By "tech-enabling", I mean layering tech into the ops where it adds value and helps to scale the business. Obvious wins are upselling, hands-free data entry to the CRM, smart traps/stations. My choice is to compete as a tech-enabled operator, rather than sell AI/SaaS to incumbents.
Similar boat here. Many of these service industries are cheap. I've built my own CRM/management system that no big company will ever touch. Even if I can sell to 1000 companies and charge them $25 a month...I'd have staff overhead, maintenance to support it. SaaS isn't some little photo editing app or something you can just launch and forget.
I'd rather grow my business and make as much money. If I can crush it with my business I'd make more than that.
Yeah agree - software needs to either do a ton more, be much cheaper, have network effects (such as connecting supply and demand), or some data benefit to avoid being built in-house or replicated.
Also for me there's an element of picking the pain I want to solve for. I've run a software company before, and prefer the tech-enabled route personally.
I agree a lightweight franchise would be attractive, though I don't like most franchising options due to the fees and lack of equity build up for the operator.
Some franchising platforms (window cleaning is a good example) don't offer much beyond sales and marketing support and some nicely designed uniforms. There's not much to window cleaning other than basic equipment, so a person's route can easily be disrupted by a new entrant who doesn't have the franchise rake to contend with.
There's a model between employment, ownership and franchising that will probably emerge as sales, marketing, ops gets easier technically.
I love this, the perfect antidote to all the stupid startup-bro grind bullshit posts.
You put in real work to understand the business landscape and typical pain points. With AI, implementing solutions has become much easier but knowing what the problems are and how to solve them hasn't.
also starting in a blue-collar field soon as an operator-ish in a facility management company. I've already lined up an awesome new SaaS in the main industry. Pest control will be one of the verticals the company has customers fo so I will be keeping an eye for it, was thinking of just starting a pest control business it self.
Does your software do anything fancy or is mostly for organization, good workflow, and being the central source of truth?
Did it require a lot of development after getting a few customers on boarded?
How long was the employment at the pest company? At any point, did anyone treat you like you were stealing their business? I thought about this approach, but I chickened out many times because of the possible confrontation.
Assuming everyone knows your acronyms is just not a good writing style.
Since I couldn't understand how s/w was going to get opossums out of anyone's basement, I think the correct decision was made: hands on!
You deserve accolades for making this choice. Good Job!
Like any physical trade, this is by it's nature a local only endeavor. So a web presence that is primarily visible to geographically local potential customers would be most effective.
Any aggregation is really just a way to skim some of the profits from the people actually doing the job. That is to say, GTM according to my definition above.
Personally, when I can't get an in-real-life personal referral to some trade, and I'm forced to do web search, I always spend extra time to try to find a web page that is put up by a local company, not an aggregator.
Things like plumers.com (this is a totally made up example, not referring to any real website) I find to be extremely irritating. Since they have absolutely nothing to do with whoever will eventually show up and do the work.
This form of aggregation through, is extremely common today, and a very large part of why the modern internet sucks.
craigslist.com (the actual website) used to be a good example of referring local services, until it was overrun with spammers and scammers.
Will this correct? Will we proceed to the dead internet? Who knows! What next weeks exciting episode to find out...
For residential / consumer markets, referrals are the gold standard and I agree to an extent about the local focus. A lot of PE (private equity) backed roll-ups result in a worse customer and worker experience as they try to force scale too fast.
Some PE companies will open a local market by initiating acquisition conversations with all local players, low ball everyone, buy some and for a short period dramatically reduce pricing to force the hold-out cohort to sell at an even lower price. Not good for communities.
The unlock to balancing scale and customer / worker experience is creating the right incentives for people to adopt the behaviors you're after. This is why bolting on SaaS or AI to established companies is tough, as the staff often don't want to change and will leave - which is bad in a tight labor market.
Searching for home services online is totally broken and is a tax on buyers and operators. HVAC contractors pay on average $600 for a closed lead from online ads, and close about one in four / one in five leads.
If you read his post, there's significant effort not "catching opossums" but waiting or churning through admin overhead - wasted time, which maybe he can translate into $. This much inefficiency is...common in many businesses.
> When I told my manager I was leaving, he said I should start my own company and give him a call when I do. So that's what I'm doing.
I love hearing stories like this, because it shows a way to be a builder without the "venture or nothing" narrative that has pervaded the tech space since the dotcom days.
It is very difficult to make a venture-backed services firm (providing services, not software) that can be immediately profitable, grow sustainably, and outperform competitors with in-house technology that's built for real on-the-ground stakeholders... at a speed that will satisfy venture investors.
But it is more possible than ever ([0]), to do this (in-house tech and all) on a bootstrapped basis - since AI reduces the engineering staff required to build, adapt, and maintain an agile best-in-class solution at single-tenant/single-customer scale. The outcome is at the least a lifestyle business, but with upside that can take the form of anything from franchising to licensing to full-fledged SaaS in the future.
I wish OOP the best of luck, and hope he's found a passion. He could go far with this approach if he ends up following through.
([0] This is not to say there are no barriers to entry. There's privilege in the word "founder," and this is no exception. And the K-shaped economy has left many brilliant would-be founders behind. But at least some barriers are lower than they once were, and that's worth appreciating.)
It will compound over time if the basics are done right (which is harder to do than I thought before this experiment)
In my previous company, we founded it with the outcome first - "take over the world" or bust. This time I think the base case is a good company, and the ceiling is the best in the industry.
He used his funding to rent four apartments in San Francisco, which he then sublet, personally, through Airbnb.
I'd really love to read a dedicated article on this side project.
Apparently, Karpathy is into AI based education business with Eureka Labs [1].
[1] Introducing Eureka Labs:
https://eurekalabs.ai/
My longterm vision is to be the agent platform for traditional industries, bridging the gap between knowledge work and physical work.
Consider being a platform coop with regional operators as members. See https://platform.coop/
The operating license holder is also on the hook for legal action if (when) things go wrong.
"Control" is interesting and I have found in all trades that people value their freedom. The good companies don't monitor employees too tightly, and are rewarded with loyalty and longer tenures generally. Of course you have to run a good recruitment and referral process to find the good people!
I’m guilty of this type of thinking and occasionally get reminded when I’m way out of my lane.
I then nearly die of internal cringe.
Reminder to myself to pick an industry that's always gonna have demand. We recently paid ~$200 for a 30 minute visit to seal off like 3 tiny holes around the perimiter of our house because of mice (actual cost of materials ~$5).
Doesn't seem like it can be a tulip if it encompasses all productive endeavors.
Think of Saudi Arabia, Iran, Putin's Russia, or Norway. I.e. risk for highly nepotic dictatorships, with the potential that it might end up well despite the odds (Norway).
Before, if you made a product that improved the lives of everyone, say you invented Google or Heinz ketchup, you could make a lot of money through that, and you did a good deed and became rich the same time. The masses of humans would reward you for delivering the benefits of your invention to them by giving you a piece of their work output.
As their work becomes less and less worth, why focus on those humans though? I am asking rhetorically of course.
An economy that thrives from innovation enriches the innovators, making them powerful. A brute in power causes the innovators to leave or in the worst case, he mass-executes them outright (think of what Stalin did in Russia). With AI, you can have a brute in power though, as an oil rig or datacenter can be protected by a bunch of machine guns.
An economy with AI everywhere will be, after a short and very innovative period, just be about who controls which resource, i.e. water for a datacenter, production lines for robots, mining rights, operational control of robot fleets, etc.
The working 95% will probably experience a sharp decrease in purchasing power, making a lot of products unaffordable to them, so consumption wise we'll have a further shift towards plutonomics. The owning top 10% will probably be affected by this major shift in consumption as well, E.g. a tower full of condos becomes worthless if the tenants can't pay rent because they got laid off, etc.
Need for robots and AI will further increase. Eventually most economic activity will revolve around those robots. It's a bit like paperclip optimizer here, whether those robots protect gay luxury space communism from counterrevolutionaries, or they project the will of the Davos council of Forbes 400, economically it will be quite similar.
There will still be human societies, humans will still talk to other humans. We won't be all exclusively conversing with LLMs, I doubt that. There will still be social mobility but it will revolve around nepotism, lying, and various escalation steps of war.
We might end up in different scenarios depending on the country, but some countries like Germany might lose relevance as most of their value lies in stuff that is going to be replaced by AI, i.e. they have less natural resources, or they have been depleted already.
We might also see companies that automate everything from end to end, from mining to producing and running weaponized robot fleets. Shareholders of those companies will do great too, if the leadership of the companies respects minority shareholder rights that is (why should they though, they will outgun any law enforcement).
Do I like this future? I don't think so. We will probably have solved cancer, communicable diseases, and aging in the next 30 years if AI continues its successful trajectory, but not sure if it will be accessible to 8 billion humans.
I'd rather grow my business and make as much money. If I can crush it with my business I'd make more than that.
Also for me there's an element of picking the pain I want to solve for. I've run a software company before, and prefer the tech-enabled route personally.
Some franchising platforms (window cleaning is a good example) don't offer much beyond sales and marketing support and some nicely designed uniforms. There's not much to window cleaning other than basic equipment, so a person's route can easily be disrupted by a new entrant who doesn't have the franchise rake to contend with.
There's a model between employment, ownership and franchising that will probably emerge as sales, marketing, ops gets easier technically.
You put in real work to understand the business landscape and typical pain points. With AI, implementing solutions has become much easier but knowing what the problems are and how to solve them hasn't.
There are lots of antiquated operators not having newer technology for pest control, which makes this area lucrative for even $50K MRR.
Go for it!
Does your software do anything fancy or is mostly for organization, good workflow, and being the central source of truth?
Did it require a lot of development after getting a few customers on boarded?
are you a 1 man show?
Assuming everyone knows your acronyms is just not a good writing style.
Since I couldn't understand how s/w was going to get opossums out of anyone's basement, I think the correct decision was made: hands on!
You deserve accolades for making this choice. Good Job!
Like any physical trade, this is by it's nature a local only endeavor. So a web presence that is primarily visible to geographically local potential customers would be most effective.
Any aggregation is really just a way to skim some of the profits from the people actually doing the job. That is to say, GTM according to my definition above.
Personally, when I can't get an in-real-life personal referral to some trade, and I'm forced to do web search, I always spend extra time to try to find a web page that is put up by a local company, not an aggregator.
Things like plumers.com (this is a totally made up example, not referring to any real website) I find to be extremely irritating. Since they have absolutely nothing to do with whoever will eventually show up and do the work.
This form of aggregation through, is extremely common today, and a very large part of why the modern internet sucks.
craigslist.com (the actual website) used to be a good example of referring local services, until it was overrun with spammers and scammers.
Will this correct? Will we proceed to the dead internet? Who knows! What next weeks exciting episode to find out...
For residential / consumer markets, referrals are the gold standard and I agree to an extent about the local focus. A lot of PE (private equity) backed roll-ups result in a worse customer and worker experience as they try to force scale too fast.
Some PE companies will open a local market by initiating acquisition conversations with all local players, low ball everyone, buy some and for a short period dramatically reduce pricing to force the hold-out cohort to sell at an even lower price. Not good for communities.
The unlock to balancing scale and customer / worker experience is creating the right incentives for people to adopt the behaviors you're after. This is why bolting on SaaS or AI to established companies is tough, as the staff often don't want to change and will leave - which is bad in a tight labor market.
Searching for home services online is totally broken and is a tax on buyers and operators. HVAC contractors pay on average $600 for a closed lead from online ads, and close about one in four / one in five leads.
GTM is ubiquitous on the business side.
If you read his post, there's significant effort not "catching opossums" but waiting or churning through admin overhead - wasted time, which maybe he can translate into $. This much inefficiency is...common in many businesses.
>when I was leaving my boss told me I should start my own company.
Genuinely or sarcastically?