Hate it when an article spends 90% of the time posing a question and 10% attempting a bare-bones answer. I got to the bottom and realized this was just an article to drive inbound deal flow.
As an AI startup founder, I worry pieces like this miss the point — we are so early into this new era of computing that it’s premature to call any winners.
It’s also simplistic, because AI-related innovation is happening across companies of every size at every layer of the tech stack. Even Open AI with its state of the art foundation models needs to compete with both open source and large companies coming out with new LLMs.
Pieces like this may also discourage bright people from taking risks and building cool things, because they worry that this is just a big company game. It’s just too early to think like that.
How about building an AI/LLM without gimping it afterwards? It's a bit tiring to see the usual "pornography is strictly forbidden" or the "As a language model, I can't <do anything even mildly outside the mainstream thought>...."-type nonsense.
Startups typically build a product that a tiny portion of users find amazing. Incumbents decide it’s not worth expending the resources to directly compete - the opportunity isn’t big enough.
Right now adopting AI is incredibly easy for the incumbents and it directly improves their core product. It’s tough to build a startup in that environment.
In my opinion the startups that will win need to stay incredibly lean or focus on fundamentally new industries.
Or look sexy enough for long enough to get aquihired.
These articles make me question what "success" means in the AI space. Is "success" just making a lot of money or raising a giant funding round or achieving a valuation? If so, then yeah, there's probably some room for at least a few startups to get there by perfectly the surfing the hype wave. I mean, so many are trying, at least one is bound to succeed by pure chance, even if nothing they're doing is particularly groundbreaking... i.e. the "wrapper around ChatGPT" business model that happens to get sold to just the right executives or PE guys or whatever.
However, if by "success" we mean truly revolutionary technology, I find it difficult to think that this will actually come out of any company trying to be profitable in the current environment. There's just too much urgency around things that are concrete, have low uncertainty, and immediate payoff, either in terms of checked boxes engineering managers can report upwards or 'initiatives' that executives can show their boards. I can't imagine doing anything resembling 'research' in that kind of environment. We'll probably have to wait until after the next recession for companies to get comfortable with throwing money and headcount at these kinds of things.
14 comments
[ 2.1 ms ] story [ 46.6 ms ] threadThis points seems to advocate for existing biz not startups
It’s also simplistic, because AI-related innovation is happening across companies of every size at every layer of the tech stack. Even Open AI with its state of the art foundation models needs to compete with both open source and large companies coming out with new LLMs.
Pieces like this may also discourage bright people from taking risks and building cool things, because they worry that this is just a big company game. It’s just too early to think like that.
I'm sure that would give them a market edge. :)
I really think that people would use it for bad, and people would at least try to take legal action against those creating/hosting the model.
Not that they would lose, but it seems like there would be legal annoyances to deal with at the very least if they don't filter their outputs.
Right now adopting AI is incredibly easy for the incumbents and it directly improves their core product. It’s tough to build a startup in that environment.
In my opinion the startups that will win need to stay incredibly lean or focus on fundamentally new industries.
Or look sexy enough for long enough to get aquihired.
However, if by "success" we mean truly revolutionary technology, I find it difficult to think that this will actually come out of any company trying to be profitable in the current environment. There's just too much urgency around things that are concrete, have low uncertainty, and immediate payoff, either in terms of checked boxes engineering managers can report upwards or 'initiatives' that executives can show their boards. I can't imagine doing anything resembling 'research' in that kind of environment. We'll probably have to wait until after the next recession for companies to get comfortable with throwing money and headcount at these kinds of things.