Many companies don’t get to Series A and very few companies get to Series B. Even if they do get to Series A or B, they won’t be able to raise the amounts you see in the news and have heavy dilution.
Very few founders have double digits percent ownership by Series B and Series C.
Liquidity of $400k or more is a lot and isn’t available for many founders.
All of this after 7 to 10 years of working 80+ hours week, no social life, loosing family, sacrificing health, taking less than $100k/year salary, constant worry of failure, dealing with ups and downs of employees, being a support system of everyone in the company while not being one for their own families, and no guarantee of success. All of this for seeing their dream come true because failure would be worse.
I think the OP should work on his company for more than 4 months and have more than 10 employees for at least a year to truly understand what it is to be a founder.
Also 20% option pool and exercising options up to 10 years are not uncommon.
> All of this after 7 to 10 years of working 80+ hours week, no social life, loosing family, sacrificing health, taking less than $100k/year salary
If you are taking less than $100k/year salary for 7 to 10 years while also absolutely no-lifing then that’s on you.
It’s true that early on you prob take ramen salary, but that’s for one or two years. You can prob scale to 200k by year 3 if your thing is viable. No-lifing when your startup is in year 5 is just a personal choice. If by year 5 you aren’t on a path of unicorn then prob it’s time to evaluate if it’s worth so much sacrifice or if you should run it as a lifestyle business (or just go do something else).
> I think the OP should work on his company for more than 4 months and have more than 10 employees for at least a year to truly understand what it is to be a founder.
Have you been an employee in a startup? Because in my experience it has a lot of the downs of the founder, but none of the ups.
> Have you been an employee in a startup? Because in my experience it has a lot of the downs of the founder, but none of the ups.
Have you been a founder? If not, I'm not sure you fully realize what goes into the job. Everyone wants to be a founder, but nobody wants to _be_ a founder.
> I'm not sure you fully realize what goes into the job.
Can it be a lot worse than working as many hours as possible and burning out? Because startup employees do that, without the compensation the founders get.
My advice would be not to do that. Set firm boundaries when you discuss the role and then enforce them. No employee will ever care as deeply about a company as its founder, and good founders understand that.
To be fair, most startups fail, and the founders of these companies can end up with similar or worse compensation than their employees. Maybe they've volunteered to take a lower salary than their early employee. Maybe because by the time they've started hiring employees, they've been working without any salary at all, burning through their savings and credit cards for a year or more before getting any meaningful funding.
Maybe I should, so that I could abuse from the employees and then explain how I deserve to get rich if MY startup succeeds but my employees don't (because it is MY startup, you see? I don't need them).
Where I come from, that's an ultra-liberal point of view. "Instead of saying that Elon Musk does not deserve 68b as a salary (because no human does), then maybe you should become ultra-rich yourself".
I mean, you could certainly start your own company, and then be more generous with your employees around these sorts of things. Sadly, you might have more trouble attracting investment, but you could probably still pull it off.
I often think about how if more people understood the median cap table life cycle from Seed to Acquisition/Shut-down/IPO, there'd be half as many VC-funded companies and twice as many bootstrapped companies every year. Thank you for sharing your experience towards that goal.
Unless you're doing some niche b2b thing where you have no personal connections (in which case, why are you doing it at all?), the differential financial returns of going with VC are often negative, if not neutral. The main diff is you can "fail up" into the investor class if you prove your worth but the business goes sideways. But even that is a dissatisfying career for most founder-type people.
To whoever needs to read this: start your own company, avoid raising money.
I think a part of the problem is that if you've chosen a market where VCs do want to invest, and you decide not to take their money, someone else is going to take it, build and grow faster than you, and out-compete you into the ground.
Sure, maybe their longer-term trajectory is unsustainable growth and disappointing surprises for founders and employees, but by that point your bootstrapped company has already shut down.
But, by all means, find a market where there's scant VC money to be found, and you can probably bootstrap for quite some time without funding. And maybe you will eventually decide to take on funding, but instead of giving 60% of your company away to get it, you only have to give away 30%. Or you decide that giving away 60% is fine, in return for 10x as much investment as you might otherwise get at an earlier stage.
I know a non-zero number of people who have gone that route, and it's worked for them. If I were to start a company, I'd aim for this model myself. But I would have to be very careful choosing my product and market.
> I think a part of the problem is that if you've chosen a market where VCs do want to invest, and you decide not to take their money, someone else is going to take it, build and grow faster than you, and out-compete you into the ground.
This is in the talking points for the VC value prop, but to be honest when you get to the bottom of all the qualifiers and explore all the examples in depth, it's a flimsy defense.
Of course you're not going to bootstrap a company with large, up-front capital requirements. That removes the risk factor "choose a market where (smart) VCs invest". It means you're fishing in $100mm up to maybe $1B markets.
Now you're left competing against the (dumb) VCs who are spinning their wheels trying to win in a market where capital doesn't actually help you grow.
That means all you have to do is survive and grow YoY – which is the default state of a sensibly-run company – until the VC-funded people give up and move on, (which they are contractually bound to do within 10 years). And even if they stick around, there are very few markets that are winner-take-all.
I think sometimes we fall prey to the mentality of thinking that things are harder than they are. The investor-funded universe completely dominates tech media, so it's perhaps not surprising. But yeah, if you think critically about each of these steps, we aren't as dependent on them as meets the eye. 100x more the case if you have good technical and business skills on your founding team.
Er... that is essentially the entire premise of the article, so I'm not sure how you can make that assertion.
To be a little more generous, the author is perhaps not saying it's sketchy, but is at least saying it's odd and unnecessary to keep this knowledge from employees.
After the Series B for my last company the three founders owned something like 45% of the outstanding shares, and when they sold took out something like 40% of the price. What were the rounds like that led to less than 10% after 3ish rounds?
> exercising options up to 10 years are not uncommon
10 year expiration is the standard, yes, but only if you stay with the company. Most still kill your options 90 days after you quit or are laid off or fired. There's been a small but noticeable trend of companies not pulling this garbage, including the article author.
> I think the OP should work on his company for more than 4 months and have more than 10 employees
Yeah, I thought it was funny that the author seemed to be speaking so authoritatively after so little experience as a founder. I've never been a founder myself, but I would put much more weight behind the words of founders who have been doing it for years and decades.
LLMs are useful to a certain extent, but from my usage they are not ready for anything harder than very basic tasks.
I feel like this megaphone about AI safety and creating a sense of doom is a strategy to increase importance of OpenAI and exaggerate the capabilities of LLMs. This era of “AI” is all about pretending that machines can think and the people working in these machines are prophets.
The call for AI safety has existed since before we broke through the Turing test with LLMs. And I personally wouldn’t call things like code generation or content-generated learning experiences for advanced topics “basic”. Not to mention where we’re headed with multimodal integration.
Many have argued for safety for decades. They’ve predicted and built the AI trajectory, they’ve been right, and we should listen.
> If one accepts that the impact of truly intelligent machines is likely to be profound, and that there is at least a small probability of this happening in the foreseeable future, it is only prudent to try to prepare for this in advance. If we wait until it seems very likely that intelligent machines will soon appear, it will be too late to thoroughly discuss and contemplate the issues involved.
~ Co-Founder of Deepmind, 2008
https://www.vetta.org/documents/Machine_Super_Intelligence.p...
Just the other day there was a double-blind study that showed a 50-50 success rate in guessing whether you were interacting with a person or GPT. That’s a turning test pass, no?
then it wasn't a canonical Turing test. The preprint accurately describes and analyzes their (indefensibly bad) experiment, but the popular press has mischaracterized it.
The canonical test gives the interrogator two witnesses, one human and one machine, and asks them to judge which witness is human. The interrogator knows that exactly one witness is human. In that test, a 50% chance of a right answer means the machine is indistinguishable from human. (Turing actually proposed a lower pass threshold, perhaps for statistical convenience.)
But that study gave the interrogator one witness, and asked them to judge whether it was human. The interrogator wasn't told anything about the prior probability that their witness was human. The probabilities that a real human is judged human and that GPT-4 is judged human sum to >100%, since nothing stops that since it's not a binary comparison. So 50% has no particular meaning. The result is effectively impossible to interpret, since it's a function both of the witness's performance and of whatever assumption the interrogator makes about the unspecified prior.
I a 5 minute casual conversation.
Also the statistics between human and AI were different in some regard (like 48% vs 56% for some quantity), I dont recall details.
Look the Turing test is very different depending on the details, and I think a lame 5min Turing test that doesnt really measure anything of i terest is a wirse concept than a 1 day adversarial expert team test thqt can detect AGI.
That linked study doesn't particularly resemble Turing's test, though? The authors asked an LLM some questions (like personality tests, or econ games), then reduced the responses to low-dimensional aggregates (like into "Big Five" personality traits), and compared those aggregates against human responses to the same questions. They found those aggregates to be indistinguishable, but that aggregation throws away almost all the information a typical human interrogator would use to judge.
Turing's interrogator also gets to ask whatever questions they think will most effectively distinguish human from machine. Everything those authors asked must appear in the training set countless times (and also corresponds closely to likely RLHF targets), making it a particularly unhelpful choice.
Turing was a WW2 era mathematician. He had no insight or understanding of intelligence, made no study of intelligent systems, and so on (he believed in ESP of all things).
Turing's test is a restatement of a now pseudoscientific behaviourism common at the time; and also, egregiously, places a dumb ape as the system which measures intelligence. If an ape can be fooled, the system is intelligent: people worshiped the sun and thought it conscious. People are desperate to analogise the world to themselves, it is a trivial thing to fool an ape on this matter.
Whatever one might make of this as a philosophical thought experiment, as a test for intelligence, its pseudoscience. What a person might, or might not believe, about a series of words sent across a wire isn't science and it isnt relevant to a discussion about the capabilities of an AI system. It is a measure, only, of how easily deceived we are.
That has nothing to do with why turing proposed it; nor does it have anything to do with general intelligence. This is just pseudoscience.
There's no scientific account of the capacities of a system with intelligence, no account of how these combine, no account of how communicative practices arise, etc. None. Any such attempt would immediately expose the "test" as ridiculous.
General intelligence arises as skillful adaptive control over one's environment, through sensory-motor concept aquistion, and so on.
It has absolutely nothing to do with whether you can emit text tokens in the right order to fool a user about whether the machine is a man or a woman (turing's actual test). Nor does it have anything to do with whether you can fool a person at all.
No machine whose goal is to fool a user about the machine's intelligence has thereby any capacities. Kinda, obviously.
Turing's test not only displays a gross lack of concern to produce any capacities of intelligence in a system; as a research goal, it's actively hostile to the production of any such capacities. Since it is trivial to fool people; this requires no intelligence at all.
> General intelligence arises as skillful adaptive control over one's environment, through sensory-motor concept aquistion, and so on.
This isn't a generally accepted definition or process.
And indeed it seems to preclude people like Stephen Hawkins who had little control over his environment (or to be pedantic, people who had similar conditions from birth).
For the purposes of my criticism of the Turing test, any discussion whatsoever about what capacities ground intelligence is already entertaining what Turing ruled out. He made the extremely pseudoscientific behaviourist assumption that no such science was required, that intelligent agents are just input-output relata on thin I/O boundaries.
Any even plausible scientific account of what capacities ground intelligence would render this view false. Whatever capacities you want to grant, no plausible ones are compatible with Turing's view nor the Turing test.
Consider imagination. You can replace a faculty to imagine with a set of models of ({prompt, reply},) histories for a human observer who is only concerned with those prompts and those replies. But as soon as anything changes in the world, you have to imagine novel things (eg., SpaceX is founded, we visit mars, a new TV show is released...). So questions such as, "what would the latest SpaceGuys TV show be like if Elon handed just launched BlahBlahRocket5 ?" cannot be given fit answers). These require the actual faculty of imagination, along with being in the world and so on.
As soon as you enter a sincere scientific attempt to characterise these features, you see immediately that whilst modelling historical frequencies of human-produced data can fool humans, it cannot impart these capacities.
> So questions such as, "what would the latest SpaceGuys TV show be like if Elon handed just launched BlahBlahRocket5 ?" cannot be given fit answers
I don't understand this at all. ChatGPT can do a great job imagining a world like this right now, and there is no substantial difference in the output of a LLM based "imagination" vs a human based "imagination".
> These require the actual faculty of imagination, along with being in the world and so on.
I think you are implying by this that human's imagination requires a consistent world model and that because LLM's don't really have this they can't be intelligent. Apologies if I have misinterpreted this!
But human imagination isn't consistent at all (as anyone who as edited a fiction story will tell you). Our creative imagination process generates wrong thoughts all the time, and then we self-critic and correct it. It's quite possible for LLMs to do this fine too!
Basically I think my point is that I believe a perfect simulation of intelligence is intelligence, whereas I suspect you don't think it is, maybe?
Yea we don't have any science of intelligence, the only thing we have is empirical data. Testing to see what works. That's why Turing tests are quite fundamental imo.
LLMs will get marginally better. But the pace of progress has already slowed down considerably, we're just seeing better usage/productization of what was already there.
also keep open that these AI researchers are as delusional as they seem. ilya sutskever has said that you can obtain any feature of intelligence by brute force modelling of text data.
It's quite possible these are profoundly naive individuals, with little understanding of the empirical basis of what they're doing.
There are (relatively simple) examples of what the transformer architecture is simply not able to do, regardless of training data, so that's simply not true.
all statistical AI systems are models of ensemble/population conditional probabilities between pairs of low-validity measures. In practice, almost all relevant distributions are time-varying, causal, and require a large number of high validity measures to capture.
eg., NLP LLMs model, eg., all books ever written using frequencies by which words co-occur at certain distances relative to other words.
But these words are about the world (, people, events, etc.) and these change daily in ways that completely change their future distribution (eg., consider what all people said about Ukraine/Russia pre/post a few hours of 2022).
The LLM has no mechanism to be sensitive to what causes this distribution shift, which can be radical for any given topic, and happen over minutes.
All models of conditional probabilities of these kinds end up producing models which are only good at predicting on-average canonical answers/predictions that are stable over long periods.
> The LLM has no mechanism to be sensitive to what causes this distribution shift, which can be radical for any given topic, and happen over minutes.
This sounds so logical and authoritative. And yet:
me> What event would cause a change in what all people said about Ukraine/Russia pre/post a few hours of 2022
GPT4O> A significant event that caused a drastic change in global discussions about Ukraine and Russia in 2022 was the Russian invasion of Ukraine, which began on February 24, 2022. This military escalation led to widespread condemnation from the international community, significant geopolitical shifts, and a surge in media coverage. Before this invasion, discussions were likely more focused on diplomatic tensions, historical conflicts, and regional stability. After the invasion, the discourse shifted to topics such as warfare, humanitarian crises, sanctions against Russia, global security, and support for Ukraine.
Right... because it's been trained on those news stories.
The point is a model whose training stopped in 2021 would not produce a history of ukraine (etc.) that a person writing in 2023 would.
The later GPTs are trained on the user-provided prompts/answers of previous GPTs, so this process (which isnt the LLM, but it's the activity of research staff at OpenAI) is what's inducing approximate tracking of some changes in meaning.
Whilst this works for any changes over-represented in the new training data, (1) the LLM isnt doing that, its the researchers; and (2) this process is vastly expensive and time-intensive; and (3) only tracks changes with a high word frequency in new data.
If you could run the months-long, 1GWh, 10s-million-USD training process each minutes of the day, you would resolve the inability of the model to track major news stores... but would not resolve its ability to track, say, the user changing their clothes.
The sensitivity to the model of stuff in the world arises because of humans preparing the training data to bring about apparent sensitivity. Absent the activity of these humans, the whole thing drifts gradually into irrelvance.
> would not resolve its ability to track, say, the user changing their clothes.
In context learning works fine for this (and does for the Russia/Ukraine change too).
But yes, sure. It can be outdated in the same way a person cut off from news can be.
We've never argued that a shipwrecked person who was unaware of news became less intelligent because of that, just that their knowledge is outdated.
Additionally, the whole point of machine learning is to make systems that learn so they remain useful.
It seems likely that a model in soon (one year? five years? one month? who knows..) will be able to continually watch video broadcast news and videos of your home, continually updating its model.
In this case it would understand both the Ukraine issue and what you are wearing. Is it now suddenly intelligent? It's true it might be more useful, but to me that is a different thing.
Thanks for uploading RCP100. Your comment is a timely one. I wanted to learn how a router works and is built and was looking for a simpler implementation.
Can you recommend any resources from which I could learn more about network programming, so that I could understand RCP100 code better?
Very few founders have double digits percent ownership by Series B and Series C.
Liquidity of $400k or more is a lot and isn’t available for many founders.
All of this after 7 to 10 years of working 80+ hours week, no social life, loosing family, sacrificing health, taking less than $100k/year salary, constant worry of failure, dealing with ups and downs of employees, being a support system of everyone in the company while not being one for their own families, and no guarantee of success. All of this for seeing their dream come true because failure would be worse.
I think the OP should work on his company for more than 4 months and have more than 10 employees for at least a year to truly understand what it is to be a founder.
Also 20% option pool and exercising options up to 10 years are not uncommon.
Source: 2nd time founder.