Note this is the intro piece to a series of articles in an Economist "Special Report" which they do from time to time on various topics. They talk about tech mo

Note this is the intro piece to a series of articles in an Economist "Special Report" which they do from time to time on various topics. They talk about tech more specifically in the "The tech antitrust paradox". However, while this not just a simple attack piece at all, and the Economist to its credit tries to consider various defenses and complexities as well, I think they miss a couple of really big ones by failing to really get down to some of the fundamental foundations tech is built on. The first issue is simple: fantastically enormous capital expenditures. The basic story here is simple and repeated, the early and medium stages of a new industry's development inevitably harvest all the low hanging fruit in terms of pure physics and math. There is no path dependency yet either. But as it matures the cost of incremental gains begins increasing non-linearly. In tech we see this most particularly in semiconductor fabrication, where the costs of the cutting edge processes that have powered everything we do have been on a J-curve for each new node. An article by by SemiEng "Big Trouble At 3nm" [1] earlier this year had a good graph from IBS show the trend in design costs alone. In terms of the fab themselves IBS estimated "3nm will cost $4 billion to $5 billion in process development, and the fab cost for 40,000 wafers per month will be $15 billion to $20 billion", and it wouldn't be surprising if that was optimistic. "7nm" (TSMC, ~Intel 10nm) has been hard and multi-billion expensive also. Economically the only way to make this work is by trying to amortize towards the thankfully very low raw material cost through enormous volume, and players must be able to front fantastic sums with very long time horizons. How does this square with tons of tiny firms? Vertical integration could be gone after so it's a pure foundry model, but that still leaves just a couple of foundries in control there. That's chips alone though, lots of other areas of electronics have the property of low resource cost combined with utterly enormous capex requirements, such as state of the art displays. Improvements in energy storage also require plants like the Gigafactory. The results are super, super important in my opinion, it's absolutely in our overall interest to keep pushing forward computation (not just silicon but whatever else becomes possible like optical computing), to reduce the price and improve the performance of energy storage technology, better human/system interfaces (retinal scanning displays next perhaps?), more reliable better memory, and on and on. We're far from "done" here. Yet this naturally takes not just long time scale dirt moving but also R&D, inevitably including some blue-sky. And blue-sky R&D in the context of tech brings us right back to government vs corps too. In principle one way to try to allow smaller firms again would be to offload a lot of this back onto government instead, which by definition has a long time horizon and shouldn't be as worried about individual short term ROI or any single area vs the long term health of the state as a whole. It can make the results of research available to all. But government R&D has not been keeping up there. Monopolies can act in effect as a pseudo-government in many ways financially, and in fact a lot of the scary parts of them can also potentially give them an escape from typical quarter-to-quarter thinking. Some of the key foundations of our current tech trace right back to places like Bell Labs and Xerox PARC, R&D outfits that performed long term research which their parents often didn't even profit from, but were themselves backed in part by the massive economic power of their parent corps. And no discussion of this would be complete without getting back to one area the government directly plays in monopolies: intellectual property. Maximalism there long since got out of control. The Economist piece doesn't give this the right level of attention IMO, particular wrt tech where rather then merely having (overlong incorrectly implemented) copyright and trademark, there are "software patents" and "business method patents" that don't just cover an implementation but basic mathematics and fundamental ideas themselves. This is a big practical part of the "moats" they talk about in their conclusion, and the difficulty small firms have in competing because that would depend on interoperability. The capex part is just really hard, and I didn't even touch on things like the space exploration SpaceX (and hopefully Blue Origin soon) has started to revolutionize after decades of doldrums. It's one thing to talk grandly about "splitting up the bad guys" and such, but asking the public to pony up tens of billions to take their place might be different (if it can't be disguised another way, like military spending maybe). And that's even assuming it'd work and government could do the job there, which it might not. R&D is more promising and the open research initiatives gaining steam now are an important part of that, but also unfortunately doesn't seem a slam dunk or very safe from politicians looking for easy things to cut and attack. "What is this even any good for anyway huh?" etc. IP is promising if only because it's purely within the power of government itself in principle, it's a creation of government in the first place anyway. Making copyright function with software like it does all previous media (source escrow for non-open/available), reducing its term to levels that reflect how modern electronic distribution has reduced costs and increased audiences along with desired churn, eliminating software and business method patents entirely, carving out far more powerful interoperability exceptions, requirements that at the least technical protection measures cannot be stacked with legal protections (or must be very term limited), would all go a long way towards increasing dynamism there. So would legal changes to cut down the EULA tape, and not just for buyers either. Producers should also be fully protected from liability by default without any contract, and contracts reserved for sufficient sophisticated entities in bespoke arrangements that wish to affirmatively bargain standard rights. It'd be good if popular anger here could be channeled towards some real positive changes, but I do fear that between lack of public or political understanding and powerful entrenched interests it will instead be channeled unproductively at best or into a real baby-and-bathwater situation at worst. To some extent big corps have even perversely taken a role defending some of our interests from governments in a way individuals cannot, would we be able to oppose intelligence agencies efforts to backdoor crypto as well as Apple for example? Of course conversely big companies have often been directly complicit and key to government authoritarian impulses and intel gathering too. It's really hard to figure out where to start, though I think IP at least is a relatively easy pickings place. ---- 1: https://semiengineering.com/big-trouble-at-3nm/ https://semiengineering.com/big-trouble-at-3nm/