I'll be honest, I really hate this guy's essays. Every single one is designed to make you feel like you're not good enough unless you're spending every day changing the world and being "high impact."
I know what I'm about to write might sound like I'm trolling but hear me out because it comes from a genuine place ...
His essays typically come across as "if you personally are not BUILDING something massive, you're wasting your time" ... though it's ironic since he himself doesn't build anything at all.
Instead, he spends his day looking for wagers to bet using other peoples money in hopes he'll produce a positive return (and keep some of the proceeds). Much like going to a casino with someone elses money.
Using his own definition of how people should spend their time, I don't understand how his own activities fall into that category.
I have one big problem with Altman which sullies my otherwise incredibly high opinion of him: he explicitly denies the orthogonality thesis and has created a culture at OpenAI that, to analogize scaling modals with nuclear energy, feels like one that knows something cool will happen when you collide large enough enriched uranium lumps together at high speed, but does not believe nuclear explosions are possible.
How he can deny the is/ought distinction when his own organization published Faulty Reward Functions in the Wild is beyond me.
And now many of those most concerned with safety, such as Paul Christiano, appear to be leaving OpenAI.
To accuse him of being incapable of building and scaling is incredibly obtuse. How much did YCombinator grow under his stewardship? I am more worried that he is too good at that.
GP says that Altman explicitly denies this theory and fosters a culture that might end up creating dangerous AI. Implied here is that there is a risk this might cause a truly catastrophic accident.
The parts about "safety" refers to acknowledging this danger and attempting to mitigate it.
I don't know enough about Altman or OpenAI to make any judgements on their approach, but that's a brief explanation.
I am not sure what the proper technique to get someone already entrenched in their perspective to broaden their possibility distribution is, though. Do you have any tips on that, as someone that appears to be engaged in the space and appears to advocate similarly?
The fact that he has an economic interest in doing so is very uninteresting -- everyone has an economic interest in their job. And when you're wealthy, the potential returns tend to be higher as well. I wouldn't confuse the fact that he stands to earn a lot of money from good bets with the fact that he's spending his time making bets that he thinks are impactful.
Alternatively, your view could be stated as "only operators BUILD things, and capital does not." That's a fine perspective, but it seems like it comes down to semantics. You need both capital and operators working in concert to make things come to reality and fruition. Imo, that's a close enough proxy to "BUILDING" but, again, semantics..
Just curious, do you consider bankers part of that equation. Because for either bootstrapped business or years ago prior to VC existed, bankers are the only channel of capital.
That doesn't have to come from a bank or a VC, but it can.
So the capital allocators who allocate to successful projects are a necessary component of the equation.
They may not "deserve" "most" of the credit, but you still need them around. So to loop back to your point -- if people like Sam have convictions about what to fund / how to choose, they can be a useful (even integral) part of BUILDING things.
Just because he's not an operator in the equation, doesn't mean that his exhortation is hypocritical. IMO
edit typos
The first is that, in the most literal sense, maybe 99.9% of the work/insight/decisions of such a capital allocator are going to be utterly useless since the creation of value will come from those rare investments that balloon out of proportion. I might be paraphrasing here as I recall a high-profile VC mentioning this exact concept. Maybe it was Altman himself?
The second is that even in an optimal case, it's up for debate how exceptional capital allocators are. If 99.9% of their qualities amount to nothing, are they really useful or actually gambling middlemen that could be excised from the process? We could imagine for the sake of argument that they do in fact contribute quite a bit and generate the insight necessary for that fateful 0.01% to happen (in similar fashion to the apocryphal tale of the specialist charging $1 for the hammer and $9,999 for having learned where to smash it), but it's unclear that a moderately energetic and highly but not superbly intelligent individual couldn't also do the same if given the opportunity to do so.
Even then, following the most charitable interpretation of the work of an allocator, is it efficient to give them the lion's share of the profits compared to those who do the actual work? I can't say I have the answers to these questions, nor am I advocating for an immediate Marxist revolution, but I am generally skeptical of the culture fostered in the tech world around this topic.
Steps to being a VC:
1. Have lots of money
2. Establish procedures that filter out the obvious losers, increasing odds of success from 1 in a million to maybe 1 in a thousand.
3. Leverage said money and procedures such that you statistically make money on that 1 in a thousand.
It's the same way someone with infinite money can keep going back to the Casino until they hit the jackpot (assuming no house limits, this is why house limits exist actually), or buy up all the lottery tickets until they find the winning ticket.
So anyone who privately funds businesses out of their own pocket is gambling, but they can't be removed because where else is the funding going to come from? Maybe you can crowd source if you have really good marketing and a popular idea. Beyond that nothing really exists on the scale of what VCs offer. If you hypothetically created a government office with a large enough budget to replicate the various VCs, perhaps it would be slightly more fair in terms of who gets access, but it would just be bureaucrats doing the gambling with taxpayer dollars instead of private entities out of their own budget, which wouldn't necessarily be any better if you look at the current state of government contracting/grants.
So sadly, their money and willingness to gamble is largely their use and why they're entitled to profits. The system isn't fair and it rewards the lucky more than the deserving, but without said system a lot of cool, revolutionary stuff would never have happened.
Even SpaceX and Tesla are indirect effects of the success of Paypal, which was initially funded by Peter Thiel.
As much as VCs would love to convince you they are oracles, the "Pros" often fare hardly better than chimps throwing darts at a board to pick investments.
I'd rather have the latter kind.
1.) Make you uncomfortable staying at your current cushy FAANG company (or any "cushy" tech job) by encouraging you to work on the weekends with your friends to write a business plan
2.) Convince all of you to leave your comfort zone and quit your day jobs
3.) Apply to accelerators like YCombinator for additional runway and advice from people who have done it
4.) YCombinator can then potentially fund your idea
And to be honest it is a win for society to create some discomfort among tech employees to dislodge them and turn them into potential founders as there is probably a significant amount of underutilized tech talent within "cushy" tech companies [1]. So its a triple win if it works out (tech employee who becomes a founder wins, YC wins, society wins).
[1] There's probably some aspect of this Google strategy that goes on at lower levels of many tech companies.
https://venturebeat.com/2015/05/09/google-has-a-secret-bench...
It felt like an impersonal scolding from a person I do not even know, a list of bullet points where you failed, or at least didn't do well enough.