Say what you want, he is a smart guy. He also has a front row seat for what OpenAi is creating.
While many question the leaks and rumors coming out of the company, Altman has a legitimate answer.
OpenAI seems to be a lot of hype. For example, they introduced the conversation ability in ChatGPT 4.o. There was also the image generating ability of Sora. Yet, as of this moment, neither is out on the market for the masses to use.
So is it hype or is there legitimacy to what is being created? This is something that we will dive into.
Moore's Law For Everything
Altman wrote a paper almost 3.5 years ago called Moore's Law For Everything. Before getting into that, consider how much things have changed since then.
This was more than 1.5 years before the release of ChatGPT 3.0. Since that time, we saw a race as other Big Tech firms started to push their solutions to market. At the time of writing, few were aware of the capabilities. Altman certainly was.
Which brings us to the present moment: is history repeating itself?
For all the speculation from the outside, perhaps Altman is, once again, reflecting his inside view. After all, he knows exactly what the next generation models are doing.
So why would he slow things down? Is it for fear of AI crushing humanity? Actually, that does not appear to be it at all.
Instead, it is giving us time to adjust. The impact of what is being created will disrupt everything. This is what the Moore's Law of Everything means.
Here is the first paragraph from that paper:
My work at OpenAI reminds me every day about the magnitude of the socioeconomic change that is coming sooner than most people believe. software that can think and learn will do more and more of the work that people now do. Even more power will shift from labor to capital. If public policy doesn’t adapt accordingly, most people will end up worse off than they are today.
Did you see that? He mentioned "the magnitude of the socioeconomic change that is coming sooner than most people believe." This is something that aligns with what many are saying. There is still a contingency that believes nothing major is happening and generative AI is nothing but hype.
Altman, back in 2021, was claiming something completely different.
He further provides his timeline:
In the next five years, computer programs that can think will read legal documents and give medical advice. In the next decade, they will do assembly-line work and maybe even become companions. And in the decades after that, they will do almost everything, including making new scientific discoveries that will expand our concept of “everything.”
Remember, we are 3.5 years into when he wrote this paper. That means we are looking at another 18 months before the major impacts are being felt.
Next Generation = Massive Change
My view is that we are going to see a major leap ahead with the next generation LLMs. What we see now is child's play compared to what we are dealing with.
Whether this is ChatGPT 5.0, Grok3, Llama4, or Claude4, the capabilities will far outpace what we are seeing today.
We know the compute is increasing. Grok3 is being trained on 100K GPUs. In contrast, Llama3.1 used 16K, the largest at the time. Of course, it not crazy to think Zuckerberg will go over 120K (perhaps even 150K) for Llama4.
It is at this point that we hit the knee of the exponential curve. The graphic above shows how this looks.
If this theory is valid, that means what we saw over the last few years is rather flat in comparison to what we will be doing. What does ChatGPT 6.0 or 7.0 look like? Perhaps Altman has insight into that also.
In the next decade, they will do assembly-line work and maybe even become companions.
Based upon the time of the estimate, that would put us in the early 2030s. That is precisely where the trends for robotics is targeting. When this is coupled with LLMs, we can see how the awareness of the world is going to grow.
A year ago I stated how Hollywood was going to be crushed by this. Here we sit and that is exactly what is happening. We are seeing the software improving, to the point where I would not be surprised to see high quality video generation by the middle of next year. This might not be feature length film quality at that point but it will be closing the gap.
Then we have finance, medical, legal, accounting, and, ultimately, factories. All of this is likely to see some form of disruption in the next 5-6 years.
Altman, if nothing else, has insight into what is going on. Even if one doesn't care for the guy, he is worth listening to.
Things are going to change greatly and few are ready.