Jim Fan tells AI community don’t worry. Be happy…and grind on code 24-7. Faster is the only way.
Many tech folks are panicking about how much DeepSeek is able to show with so little compute budget. I [Jim Fan] see it differently – with a huge smile on my face. Why are we not happy to see *improvements* in the scaling law? DeepSeek is unequivocal proof that one can produce unit intelligence gain at 10x less cost, which means we shall get 10x more powerful AI with the compute we have today and are building tomorrow. Simple math! The AI timeline just got compressed.
Here’s my 2025 New Year resolution for the community:
No more AGI/ASI urban myth spreading.
No more fearmongering.
Put our heads down and grind on code.
Open source, as much as you can.
Acceleration is the only way forward.
Whether you like it or not, the future of AI will not be canned genies controlled by a "safety panel". The future of AI is democratization. Every internet rando will run not just o1, but o8, o9 on their toaster laptop. It's the tide of history that we should surf on, not swim… pic.twitter.com/VPmgV8wR1C
— Jim Fan (@DrJimFan) January 24, 2025
Brian Wang is a Futurist Thought Leader and a popular Science blogger with 1 million readers per month. His blog Nextbigfuture.com is ranked #1 Science News Blog. It covers many disruptive technology and trends including Space, Robotics, Artificial Intelligence, Medicine, Anti-aging Biotechnology, and Nanotechnology.
Known for identifying cutting edge technologies, he is currently a Co-Founder of a startup and fundraiser for high potential early-stage companies. He is the Head of Research for Allocations for deep technology investments and an Angel Investor at Space Angels.
A frequent speaker at corporations, he has been a TEDx speaker, a Singularity University speaker and guest at numerous interviews for radio and podcasts. He is open to public speaking and advising engagements.
For the stability of democracy, open source.
For the stability of society, closed systems.
We don’t always like the medicine we need to take.
3D printing, robotics, and AI seem to be starting to converge.
I tried it on Mac, but wasn’t satisfied with the results. Next step is to load it on Juno this weekend.
I think a major issue that scared the market is the realization that the “AI technological landscape” is not well charted. Companies invested billions of dollars (either their own or VC capital) because they thought that the only way to get a competitive advantage was the size of the computing farm. While this might still be true in general terms, it might also mean that if someone develops a couple of orders of magnitude better model and it is not so kind to release it open source, you are stuck with paying 10-100 times the electric bill (and the costs of the capital investment in the servers).
The general scare is not that computing power will be useless. What scares investors is being stuck with billions of dollars of investments that have to make a profit (covering billions of dollars of costs) while a competitor might develop and run a product at a fraction of the cost.
It is not about depseek itself. It is about the realization that even the experts in the field (who are the ones who persuaded investors to invest hundreds of billions) could be completely blindsided.
> 10x less cost, which means we shall get 10x more powerful AI with the compute we have today and are building tomorrow. Simple math!
That’s not necessarily the case. At least, define “10x more powerful”.
For example, a 80 IQ person can’t solve problems that a 120 IQ person can, just investing a 50% more time. That’s not a linear scale.
If you said that we can do 10x more tasks (so the same computing per tasks) then that’s right. But 10x more “powerful”, can be misleading.
In fact, the -o3 route seems to require exponential power to get logarithmic improvements.
We will probably need new reasoning models. Ones that can abstract over abstract, and not only get better results, but better reasoning and abstraction strategies over time so it can compensate the increment of power to finally beat human thought for a fraction of the power cost.
Of course I am happy about this development. One of the blogs I follow convinced me that DeepSeek is real. I believe very strongly in the “democratization” (really decentralization) of all technology (and politics for that matter). DeepSeek put a smile on my face yesterday morning.
The very real danger of AI (as opposed to the hypothetical AI takeover) is the emissions caused by huge power requirements. There is no question we should be celebrating if it can be done more efficiently. Apart from this, advances, especially open source, should be welcomed. Competition drives innovation. As Altman has already commented “it’s legit invigorating to have a new competitor!”
10X smarter attack drones. What’s not to love???
I was just discussing with my son the possibility of integrating Deepseek into a FIrst Robotics robot, to improve autonomous mode performance.
He commented that the DOJ would be very enthusiastic if they did that, because most of the sponsorship money comes from defense contractors, and being on a team gives you a real leg up on getting a job with one of them.