The secret to success for OpenAI, Anthropic and labs is the vision that they saw 10 years back and kept working on it. We're in awe of how models like GPT-6 Astra and Claude Opus 5.5 are performing today, but it's important to understand that they've been working on this before we knew about AI.
The next big thing is Robots and some stealth company building today is going to be a trillion-dollar giant in few years time.
That’s just hindsight. People work on a lot of different things. Breakthoughs and cumulative improvements reaching a tipping point might happen or not.
Even for domestic robots people don't care, see just about every cloud connected smart vacuum robot and the numerous data leaks
And for companies, most of them are fine using SaaS and cloud based hosting platforms, so why would it be any different for robotics?
Sure there is the latency argument, so at least something doing very short term decision making needs to run on device or at the edge, but for most task the end to end time is in the single didit seconds to minuites, so the extra 100ms or whatever that you get going out to some model in the cloud isn't a deal breaker.
I assume the first versions will have local AI handling basic functions like balance and motion, and remote AI handling higher-level decisionmaking. I also agree that robots feel like something we should approach very cautiously, but of course we’re not.
This feels like a retcon of OpenAI. Their vision from back then is unrecognizable because they had to completely pivot after they saw the value of transformers. They kicked off the LLM arms race in 2022 because they thought they were competing with Google.
Like fuzzy logic before it, this arm of AI development will most likely hit a wall, find where it is useful, and then become cheap and embedded all over to the point where we won't even recognize it as AI anymore.
Right. If you're new to the area, it may seem like AI came out of nowhere in 2022. If you dig a little, you'll be amazed that it somehow came from nowhere in ~2012. If you dig even more, you realize there was a wave in the late 90s, early 2000s about "machine learning" (e.g. SVMs) and before it there was an 80s wave of both neural nets, agent models, and logic-based AI, probabilistic graphical models. Then you dig more and you realize AI originated from that Dartmouth workshop by Minksy and others in the 50s. Then you dig more and realize McCulloch and Pitts already modeled neural nets as little logic circuits in the 1940s. Then you realize the role of Shannon, Turing etc. Then you realize that computers actually arose in a milieu with a much more AI-shaped vision, cybernetics etc. than what we today think of as computing (PCs etc). And the precursors in the thought-formalization and mechanization trend in math and philosophy at the start of the 20th century. And even more back Leibniz's calculus ratiocinator and "calculemus!" slogan to settle debates by reducing argumentation to computation.
The point is, typically when something seems like it came out of nowhere, it just means you didn't dig deep enough. Ideas don't come at an instant, fully formed like Athene from Zeus' forehead. It's brick by brick, one twist on an existing idea and zeitgeist at a time.
"The people on top desperately want us gone, and when it happens it'll happen so quickly you won't know what hit you. Software engineers need to recognize this common threat and organize (labor) sooner than later. Even more important is for all engineers to plan for an imminent future where developers are not paid like they are now, if at all. We have it good, but we will be automated away like everybody else, just a little later."
Remember this when some corporate talking head talks about this stuff like it wasn’t predictable or preventable. Or when a coworker talks in a defeatist manner about it.
It's interesting to see how most of the discussion on the 'program that can write programs' was approaching it from a very traditional-AI point of view - how do we define the problem space, how do we make programming languages more amenable to it, etc. basically just envisioning a more sophisticated version of a traditional code generator. Nobody really foresaw that we would have a general intelligence that can write programs effortlessly by just dumping the entire internet into a simple algorithm.
Right? Amazing how confidently wrong one can be, by assuming there is no other way to approach a problem, other than the one your current intellectual tools point at.
A complex simulation with many long-lived agents: We’re interested in building a very large simulation with lots of different agents in it that can interact with each other, learn over a long period of time, discover language, and accomplish a rich variety of goals.
This sounds rather like the plot to Greg Egan's "Crystal Nights" (2008)
Ten years ago I didn't know AI existed, it was a term for the future. Now is the future they built for us. Thank you guys, all of you, whether people love you or hate you it is undeniable the progress you have unleashed for the benefit of mankind and history will be grateful about your legacy
>it is undeniable the progress you have unleashed for the benefit of mankind and history will be grateful about your legacy
Keep watching, then. I'm not so sure history will be quite as grateful as you seem to think. The future has a funny way of looking very different from what people expected when they were busy building it.
Once men turned their thinking over to machines in the hope that this would set them free. But that only permitted other men with machines to enslave them.
Not sure what happened on this page, but I saw so much AI tells on this particular page that I had to check archive.org and apparently the page has only been around for two years.
Although they did loosely stick to their vision they also didn't right? I mean this as a company not the extremely subsidized research. They aren't open sourcing anything, they are now for profit, they are advertising, etc.
I wouldn't view them as consistent at anything other then myopicly following the more or less obvious trends required to sustain LLM architectures over the years
To me it's more about the funders staying engaged. For the people working in the startup as long as it's exciting and you're getting paid enough it seems really fun to get that much runway - why would you leave unless there are personality/direction conflicts??
The secret to success for OpenAI, Anthropic and labs is the vision that they saw 10 years back and kept working on it. We're in awe of how models like GPT-6 Astra and Claude Opus 5.5 are performing today, but it's important to understand that they've been working on this before we knew about AI.
The next big thing is Robots and some stealth company building today is going to be a trillion-dollar giant in few years time.
That’s just hindsight. People work on a lot of different things. Breakthoughs and cumulative improvements reaching a tipping point might happen or not.
Agreed, Quantum Computing is a nice example of this.
1 reply →
> The next big thing is Robots
IMHO, robots won't become a thing until they're running local AI, at which point they will become the thing that crushes humanity.
Why?
Even for domestic robots people don't care, see just about every cloud connected smart vacuum robot and the numerous data leaks
And for companies, most of them are fine using SaaS and cloud based hosting platforms, so why would it be any different for robotics?
Sure there is the latency argument, so at least something doing very short term decision making needs to run on device or at the edge, but for most task the end to end time is in the single didit seconds to minuites, so the extra 100ms or whatever that you get going out to some model in the cloud isn't a deal breaker.
1 reply →
I assume the first versions will have local AI handling basic functions like balance and motion, and remote AI handling higher-level decisionmaking. I also agree that robots feel like something we should approach very cautiously, but of course we’re not.
1 reply →
This feels like a retcon of OpenAI. Their vision from back then is unrecognizable because they had to completely pivot after they saw the value of transformers. They kicked off the LLM arms race in 2022 because they thought they were competing with Google.
Like fuzzy logic before it, this arm of AI development will most likely hit a wall, find where it is useful, and then become cheap and embedded all over to the point where we won't even recognize it as AI anymore.
Nvidia also had a long-term vision from which they are clearly reaping the rewards.
> they've been working on this before we knew about AI
Don't confuse AI with LLMs. "We" know about AI for a long time. We even have a term for when AI fails expectations, AI winters.
Right. If you're new to the area, it may seem like AI came out of nowhere in 2022. If you dig a little, you'll be amazed that it somehow came from nowhere in ~2012. If you dig even more, you realize there was a wave in the late 90s, early 2000s about "machine learning" (e.g. SVMs) and before it there was an 80s wave of both neural nets, agent models, and logic-based AI, probabilistic graphical models. Then you dig more and you realize AI originated from that Dartmouth workshop by Minksy and others in the 50s. Then you dig more and realize McCulloch and Pitts already modeled neural nets as little logic circuits in the 1940s. Then you realize the role of Shannon, Turing etc. Then you realize that computers actually arose in a milieu with a much more AI-shaped vision, cybernetics etc. than what we today think of as computing (PCs etc). And the precursors in the thought-formalization and mechanization trend in math and philosophy at the start of the 20th century. And even more back Leibniz's calculus ratiocinator and "calculemus!" slogan to settle debates by reducing argumentation to computation.
The point is, typically when something seems like it came out of nowhere, it just means you didn't dig deep enough. Ideas don't come at an instant, fully formed like Athene from Zeus' forehead. It's brick by brick, one twist on an existing idea and zeitgeist at a time.
10 replies →
Discussion at the time: https://news.ycombinator.com/item?id=12181881 (235 points, 197 comments)
"The people on top desperately want us gone, and when it happens it'll happen so quickly you won't know what hit you. Software engineers need to recognize this common threat and organize (labor) sooner than later. Even more important is for all engineers to plan for an imminent future where developers are not paid like they are now, if at all. We have it good, but we will be automated away like everybody else, just a little later."
- superswordfish on July 28, 2016
Remember this when some corporate talking head talks about this stuff like it wasn’t predictable or preventable. Or when a coworker talks in a defeatist manner about it.
It's interesting to see how most of the discussion on the 'program that can write programs' was approaching it from a very traditional-AI point of view - how do we define the problem space, how do we make programming languages more amenable to it, etc. basically just envisioning a more sophisticated version of a traditional code generator. Nobody really foresaw that we would have a general intelligence that can write programs effortlessly by just dumping the entire internet into a simple algorithm.
Right? Amazing how confidently wrong one can be, by assuming there is no other way to approach a problem, other than the one your current intellectual tools point at.
That thread is a goldmine. Very humbling.
A complex simulation with many long-lived agents: We’re interested in building a very large simulation with lots of different agents in it that can interact with each other, learn over a long period of time, discover language, and accomplish a rich variety of goals.
This sounds rather like the plot to Greg Egan's "Crystal Nights" (2008)
https://www.gregegan.net/MISC/CRYSTAL/Crystal.html
> Build an agent to win online programming competitions. A program that can write other programs would be, for obvious reasons, very powerful.
Interesting that Anthropic led the way here.
> Authors: Ilya Sutskever, Dario Amodei, Sam Altman
Only one remains at OpenAI.
That's good! The economy is thankful for it.
The other two left because they had fundamental disagreements with sama about how to run a company and how to shepherd humanity into an AI future.
If you ask me, it'd be better if none of them remained at OpenAI.
Ten years ago I didn't know AI existed, it was a term for the future. Now is the future they built for us. Thank you guys, all of you, whether people love you or hate you it is undeniable the progress you have unleashed for the benefit of mankind and history will be grateful about your legacy
>it is undeniable the progress you have unleashed for the benefit of mankind and history will be grateful about your legacy
Keep watching, then. I'm not so sure history will be quite as grateful as you seem to think. The future has a funny way of looking very different from what people expected when they were busy building it.
Once men turned their thinking over to machines in the hope that this would set them free. But that only permitted other men with machines to enslave them.
Its really impressive that they stuck to their vision over the long term. This is difficult and more challenging than it sounds.
Not sure what happened on this page, but I saw so much AI tells on this particular page that I had to check archive.org and apparently the page has only been around for two years.
https://web.archive.org/web/20260000000000*/https://openai.c...
It's only been saved as that URL for the last two years.
Where do you think 'AI tells' came from? :-)
Although they did loosely stick to their vision they also didn't right? I mean this as a company not the extremely subsidized research. They aren't open sourcing anything, they are now for profit, they are advertising, etc.
I wouldn't view them as consistent at anything other then myopicly following the more or less obvious trends required to sustain LLM architectures over the years
To me it's more about the funders staying engaged. For the people working in the startup as long as it's exciting and you're getting paid enough it seems really fun to get that much runway - why would you leave unless there are personality/direction conflicts??
[dead]