I have a feeling that OpenAI can surpass Apple, Microsoft, Google, Meta, Amazon, etc. due to its laser focus on a technology that is truly transformative and disruptive.
I doubt it, recreating what they've done in a short span of time only requires a small amount of capital. All the big players probably have teams working on this internally, and there are probably dozens of stealth startups gathering money to compete.
I just don't see OpenAI scaling to the moon when the moat around their product is so shallow and disruptable. What is their long-term play besides having an API that can be replaced with by changing a POST request URL?
How do you define 'small amount of capital'? Assuming you have to buy the processing from one of the big 3 (likely not possible to own the hardware outright), you're looking at 10s of millions of dollars per training run.
It's certainly not outside reality for the well-known software companies to have divisions working on it, but I don't think we're quite at the point where garage startups can afford the compute.
> likely not possible to own the hardware outright
Why not? The majority of these models are running on fairly ubiquitous datacenter GPUs. They're being trained on highly available hardware. You're right that there's a compute barrier, but you get around that by throwing money at the problem. If someone was sufficiently motivated, they could probably dethrone OpenAI in 2 months with enough funding.
they've put themselves in a extremely interesting position, the only way they stay on top is through innovation. there definitely is a spot for a gigantic AI company.
It's going to take a big tech company or partnership wit a tech company to compete in this space, given the massive compute requirements.
Apple is a hardware company and it's doubtful they will ever be a direct competitor.
Google is an AI powerhouse, but they are a mess internally and seem to be too terrified of their brand reputation to actually release any of their work as products. They have potential, but might fail to execute.
Amazon seem like a strong potential if they can build on their AWS products.
Meta is off chasing VR, so probably not unless they execute a big pivot. Lots of impressive researching coming out of them though.
Microsoft already seems to be in the lead with their investment in OpenAI and the speed they are pushing it into their products.
So if the LLM potential is really as big as it seems, some combination of these companies is probably going to own the market. Whether thats an extra 10 billion of market cap or 10 trillion of market cap remains to be seen. personally I have some small investments, but it's a single digit percent of my NW.
I think the more precise question is can Sam Inc. become bigger than Elon Inc.
Both extraordinarily smart and hard working.
I really like Elon's anti-patent stance. SamA backpedaled a bit on the openness with OpenAI, but I wouldn't be surprised if that was a short term move and in the long run it will become more open again.
I think Elon is in the much better position because he sells physical objects.
I think OpenAI still has very existential risks to their business model. In the long run open source is an OOM cheaper way to create software than closed source, and as long as Apple keeps cranking out faster processors the open source community will be able to reproduce and exceed OpenAI's cloud offerings. I think SaaS may be more of a fad than people realize.
But Sam has diversified big with big bets on fusion and other things. I would expect he comes up with a business in the real world that has more of a moat than a business dependent on (c)opywrong law.
It's a fun competition to watch.
My money is definitely on Elon, but SamA is probably the next best entrepreneur in the game today.
Here we go again. I don't know if it's the PR machine gaining steam, but there has been a flurry of posts propping up Sam Altman across social media lately. He's just another serial investor who struck timely fortune during a great boom. Why do we feel the need to put financiers on a pedestal? These people, by the very nature of their role, have interests that are largely unaligned with society and history generally shows them to be a net negative on the world in the end.
Besides, an orgranisation is always greater than any individual, and we should judge organisations, or any groups of people for that matter, as their whole and not singularly. Sometimes it feels us plebians always seek out a superhero in our lives. With Musk going full villain, there is a void that needs to be filled. Perhaps, in a world where all else is lost, this is the one thing that can keep the spirit going.
> Why do we feel the need to put financiers on a pedestal?
I don't put him on a moral pedestal. Just at the game of startups, he's one of the best. I don't take my moral advice from Lebron but I respect the heck out of him.
> interests that are largely unaligned with society
I think the jury is still out on SamA.
> With Musk going full villain
I can't disagree strongly enough with this. Musk is a hero. I'm not sure people realize how much censorship was winning in the USA prior to Musk buying Twitter. Scary days.
I don't understand Googles hesitation in this space, when they thought Facebook posts were going to replace (a portion) of search they went all in on social platforms. This is clearly a much greater threat to their core business, so much so that it's allowed Microsoft to reinvigorate bing!
I've never worked at google, but I've worked for them and Meta, the data they have access to is unbelievable (and I'm sure I only saw a very small subset of it).
If they don't successfully make the paradigm shift then their dominance in search is at an end. (Especially with governments regulating search results like what's happening in Canada with bill C18)
I don't think it is hesitation. They probably just do not have results that are comparable, yet. In that case it is better to not release anything than release results that show that your model is 1/4 as good as that of your competitor.
My belief is that Google was betting more on DeepMind's vision, which is closer to "thinking" than what LLMs models do.
LLMs have demonstrated some usefulness for sure, but it's no AGI. It makes many mistakes and doubles down on them at times. If Google can nail AGI on easier taks / games / simulations / etc, they'll win that race, cause it'll be easier to add a language layer to a thinking model than to tweak a language model into thinking.
What is OpenAI's moats? Those companies are valuable because of the moats they've created (users, market position, and economies of scale).
OpenAI is currently the market leader but what is keeping them there? Their competitors have a lot of expertise and access to the compute needed.
If anything because OpenAI is more of an API provider than having their own end-users, any of their current customers could switch to another cheaper/better AI vendor without their end-users even knowing. Diminishing OpenAI's USP even further.
More importantly I wouldn't be surprised if OpenAI's top talent is poached by Google/Facebook. Or are there some contracts that explicitly forbid this? I think such contracts would be illegal no?
Probably not. I do not think they have much in terms of trade secrets. Now that they showed what can be achieved it is only a matter of time, before other teams catch up. I would be surprised if they hold their lead for more than 5 years.
I think no. OpenAI will not have GPT family models as exclusive technology. GPT technology is not accessible only to OpenAI. Other companies can make and I think will soon make alternative solutions to GPT family.
The research momentum behind ChatGPT is in the open, widely understood, and has no meaningful IP protections. They’ve now demonstrated where the market is, and money will come in and stake out many different and competing positions. Several of those will probably eclipse everything we’ve seen in these last few weeks.
Microsoft, meanwhile, has come in and begun guiding them through product development and engineering as their own effort to stake such a position. Microsoft is reasonably good at this but has no monopoly on it either.
In any case — because of the Microsoft attachment, they’re no MySpace or Altavista and won’t simply wither away, but they’re basically just a vassal whose fate and glory is at the whim of their patron.
23 comments
[ 3.2 ms ] story [ 57.7 ms ] threadI just don't see OpenAI scaling to the moon when the moat around their product is so shallow and disruptable. What is their long-term play besides having an API that can be replaced with by changing a POST request URL?
It's certainly not outside reality for the well-known software companies to have divisions working on it, but I don't think we're quite at the point where garage startups can afford the compute.
Why not? The majority of these models are running on fairly ubiquitous datacenter GPUs. They're being trained on highly available hardware. You're right that there's a compute barrier, but you get around that by throwing money at the problem. If someone was sufficiently motivated, they could probably dethrone OpenAI in 2 months with enough funding.
I'd argue Google is sufficiently motivated to dethrone OpenAI --- so why haven't they? They have plenty of money to throw at the problem as well.
Apple is a hardware company and it's doubtful they will ever be a direct competitor.
Google is an AI powerhouse, but they are a mess internally and seem to be too terrified of their brand reputation to actually release any of their work as products. They have potential, but might fail to execute.
Amazon seem like a strong potential if they can build on their AWS products.
Meta is off chasing VR, so probably not unless they execute a big pivot. Lots of impressive researching coming out of them though.
Microsoft already seems to be in the lead with their investment in OpenAI and the speed they are pushing it into their products.
So if the LLM potential is really as big as it seems, some combination of these companies is probably going to own the market. Whether thats an extra 10 billion of market cap or 10 trillion of market cap remains to be seen. personally I have some small investments, but it's a single digit percent of my NW.
Both extraordinarily smart and hard working.
I really like Elon's anti-patent stance. SamA backpedaled a bit on the openness with OpenAI, but I wouldn't be surprised if that was a short term move and in the long run it will become more open again.
I think Elon is in the much better position because he sells physical objects.
I think OpenAI still has very existential risks to their business model. In the long run open source is an OOM cheaper way to create software than closed source, and as long as Apple keeps cranking out faster processors the open source community will be able to reproduce and exceed OpenAI's cloud offerings. I think SaaS may be more of a fad than people realize.
But Sam has diversified big with big bets on fusion and other things. I would expect he comes up with a business in the real world that has more of a moat than a business dependent on (c)opywrong law.
It's a fun competition to watch.
My money is definitely on Elon, but SamA is probably the next best entrepreneur in the game today.
Naive people is why such strategies works.
Besides, an orgranisation is always greater than any individual, and we should judge organisations, or any groups of people for that matter, as their whole and not singularly. Sometimes it feels us plebians always seek out a superhero in our lives. With Musk going full villain, there is a void that needs to be filled. Perhaps, in a world where all else is lost, this is the one thing that can keep the spirit going.
I don't put him on a moral pedestal. Just at the game of startups, he's one of the best. I don't take my moral advice from Lebron but I respect the heck out of him.
> interests that are largely unaligned with society
I think the jury is still out on SamA.
> With Musk going full villain
I can't disagree strongly enough with this. Musk is a hero. I'm not sure people realize how much censorship was winning in the USA prior to Musk buying Twitter. Scary days.
He’s the originator of the Worldcoin scam: https://www.buzzfeednews.com/article/richardnieva/worldcoin-...
Insane.
I've never worked at google, but I've worked for them and Meta, the data they have access to is unbelievable (and I'm sure I only saw a very small subset of it).
If they don't successfully make the paradigm shift then their dominance in search is at an end. (Especially with governments regulating search results like what's happening in Canada with bill C18)
LLMs have demonstrated some usefulness for sure, but it's no AGI. It makes many mistakes and doubles down on them at times. If Google can nail AGI on easier taks / games / simulations / etc, they'll win that race, cause it'll be easier to add a language layer to a thinking model than to tweak a language model into thinking.
OpenAI is currently the market leader but what is keeping them there? Their competitors have a lot of expertise and access to the compute needed.
If anything because OpenAI is more of an API provider than having their own end-users, any of their current customers could switch to another cheaper/better AI vendor without their end-users even knowing. Diminishing OpenAI's USP even further.
The research momentum behind ChatGPT is in the open, widely understood, and has no meaningful IP protections. They’ve now demonstrated where the market is, and money will come in and stake out many different and competing positions. Several of those will probably eclipse everything we’ve seen in these last few weeks.
Microsoft, meanwhile, has come in and begun guiding them through product development and engineering as their own effort to stake such a position. Microsoft is reasonably good at this but has no monopoly on it either.
In any case — because of the Microsoft attachment, they’re no MySpace or Altavista and won’t simply wither away, but they’re basically just a vassal whose fate and glory is at the whim of their patron.