A non-obvious second effect of the enormous AI capex spend may be that it's masking the deterioration of the US economy under the surface. While the aggregate economic stats may look just fine, economy is becoming ever more K-shaped with the lower half of th economy struggling ever more in terms of keeping up with cost of living increases, young graduates struggling to find jobs, and credit card and auto loan bills pilling up.
The seemingly non-negotiable enormous capex spending has also an underappreciated crowding out effect on the rest of the real economy. First, the AI capex drives up input cost for non-AI manufacturing industries from materials to labor. Indeed, manufacturing has been shrinking while data center capex surges.
Secondly, if companies have to spend more on capex and expect more productivity from AI, they’ll hold back on hiring. It has already eaten into the hiring plans of even the highly profitable large cap tech companies. To rephrase @pmarca, AI is eating the world, but maybe not quite in the way we expected.
Miyazaki spent his entire life building one of the most expansive and imaginative bodies of work, all so you could rip it off and use it as a filter for your vacation photos.
Not into this one bit. Protect artists.
Lot of talk about Arc Browser “cleaning up the internet for you.” If AIs (Bing chat included) read our blogs, skip ads, and generate summaries, there’s no incentive for bloggers or writers to write. The AIs will quickly end up summarizing another AIs confidently generated crap.
I still haven't heard a good answer to this question, on or off the podcast.
AI researchers often tell me, "Don't worry bout it, scale solves this."
But what is the rebuttal to someone who argues that this indicates a fundamental limitation?
@fkadev The modern web isn't an innovation medium anymore, but controlled by handful of companies like Google and Microsoft.
Even Linux is commercialized successfully.
@fkadev The nature of open source changed nature since the 90s. I don't think it's possible to bring back that culture unless there's a significant change in the control of technology.
@fkadev Şirketler ihtiyaçlarını tabi ki daha ucuza istihdam etmeye çalışacaklar. 2010'da iOS için yazılım geliştirtmek için servet ödeniyordu.
Ana problem bence Türk yazılım piyasasının büyüyememesi.
@fkadev Yazılım 90'lardaki gibi "high cost of entry" bir alan değil artık. Elbette ki piyasaya göre daha iyi ödeyen bir alan daha çok bireyi, şirketi kendine çekecek.
Türkiye'de 453 bin müteahhit varken bu sayı Almanya'da 3800'mış mesela.