Agriculture is much slower, every iteration may be is a year, or (in tropical climates) half a year. Microelectronics is comparably slow, and even more unforgiving about making mistakes. Building robots does not scale ls easily as producing chips, let alone software.
These areas need a different model of investment, with a longer horizon, slower growth, less influence of fads, better understanding of fundamentals. In some areas, DARPA provided such investment, with a good rate of success.
If you have money, the returns you'll get elsewhere are much less attractive, and can only be justified if they're very safe investments.
To fund a similar sized hardware start-up you need a full lab andddd already the proposition is dead.
Contrast this with biotech venture capital which has been doing well for decades, often investing more capital in a year than software VCs. The difference is that all the research, clinical trial, and manufacturing expertise is already here and concentrated in a few localities like South San Francisco, San Diego, and Boston.
I feel like anything relevantly practical is denied investment.
But when it comes to anything flashy and hip, a train of dump trucks filled with cash couldn’t deliver money as quick as the VC dollars that flow into to startups with no business model and no hope of being profitable…
Yeah, I get that startups should invest profits and not actually make profits for a while… But when they’re on their 4th round of funding with thousands of employees… shouldn’t they at least try to be a bit more financially responsible?
They don't know anything else.
https://www.currentmarketvaluation.com/models/s&p500-mean-re...
- The chatbot people have a personal attachment to
- The processing tool.
In the second, you only care about the result. Something like Claude Code can call any other provider if that's cheaper and visa-verse. Once I have the result, my dependency / lock-in is no more than a brand of toilet paper. The providers will have to do the 'capitalism thing' and compete.
It's almost like WeWork's, valuated at IT levels by being in the style, only for investors to eventually figure out the marginal production costs are not reducible to near 0, and you can't just bully out competition / network-effect to get a monopoly.
And this applies to any company that wraps and re-sells AI.
Something the tech-VC world is so unfamiliar with, it's scrambling to present the truth of what is 'econ-101' for the rest of the world.
This YCW18 ag company was acquired less than 3 years in by John Deere for $250M: https://techcrunch.com/2021/08/05/john-deere-buys-autonomous...
That said, allowing VC into 401ks and such I would agree is an abominable idea, because this stuff isn't marked properly until it is in distress. Actually, that area could use better regulation. Volatility laundering is already a systemic risk. Many of these vehicles have creative ways to not mark to their market value, which makes pension fund managers and leered participants happy because it greatly improves the perceived risk metrics and performance, at the equal expense of cloaked fragility.
But perhaps just let them have a thunderdome, and if they want to breach the walls and enter areas like retirement funds where society agrees standard are higher, there is a strong set of filters/regulations that must be adhered to.
(Also, in neither country is the majority of its GDP comprised of websites.)
The number of countries producing leading edge semiconductors is actually small. The number of companies doing this work is very small, too. Although much of the economy needs chips to operate, those leading chips are concentrated in the production of a very small number of companies.
It's slightly weird to me how foreigners seem to look on the Trump era as personifying the US to a greater degree than e.g. the Biden or Obama eras. Trump is not especially popular right now: https://www.economist.com/interactive/trump-approval-tracker
Is that the particular one you're referring to?
