I have a lot of respect for the author, Tim Berners-Lee. It’s not that long ago I gave a talk on the HTTP, the underlying invention of his that made the WWW possible; and many in high-tech won’t have the careers they have absent Berners-Lee’ invention.
Tim Berners-Lee came from a time when low cost impactful innovation was possible in the field.
As he said himself, he wrote the code for WWW on a single computer in a small room by himself. Much of the technologies that underpin today’s IT infrastructure were invented by very small teams with shoestring budgets. But this is much harder to do now. The low hanging fruits have mostly been plucked. Now the AI arms race requires multiple $billions for credible admission. I don’t see quantum computing being cheaper than AI. It’s easy for a small team to give away the fruits of a few person-years of (part-time) work, but a lot harder for a multi-$billion corporation to give away the fruits of a multi-$billion investment.
Meta did the AI world a lot of good open sourcing LLaMA, but that’s an exception rather than the norm. I don’t see many other companies open sourcing their AI models. Even Zuckerberg indicated that Meta will be more selective about what parts/features of future AI models to open source. The future for this vision about freely sharing IT innovations with the world does not look that bright at the moment.
