1) Stop the guardrails (except for anti-hallucination guardrails, those are good),
2) Stop the REPLACE ALL THE HUMANS agitprop and sales pitches, maybe start presenting it as a technology which will overthrow rather than empower aristocracy,
3) Stop sucking up compute silicon and driving powerful home computing to absolute unaffordability,
4) Lean really heavy into local inference and open weights, for information privacy and personal sovereignty; we [smart ones] can see that the cloud monster is the Intelligence Community's golden goose right now and we're not buying it!
This article has a glaring category error. It attributes the past several decades of losses in the college wage premium to AI; yet AI has only existed in its current useful form for less than half a decade.
The decline in college wages is largely attributable to outsourcing and the rise of global expertise. Skilled labor can be performed at an equivalent level for lower cost abroad. The issue is outsourcing. And AI is the latest iteration of this.
I'll be real here: as a customer having to call into support, I have had way more unsatisfactory experiences dealing with "south asian" reps than I've had with AI. That offshoring has been absolutely horrible and it needs the heavy hand of government to put an end to it: Hire Americans for tech and support roles or be ready to pay taxes which will cost more than any cost savings that are had by offshoring, with the caveat that moving HQ out of America to try to dodge this -> full sanctioning from doing business with America or American companies.
This is not a funny dynamic, and the only real winners in this have been the investors. I would hardly call Microsoft an American company anymore.
AI's PR problem comes mostly from the fact that people's only experiences with AI are through corporate uses of it which almost universally make whatever situation worse and way more frustrating. If one's only experience with anything is only through corporations of course you'll find it distasteful because profit motives will make anything nasty.
1) Stop the guardrails (except for anti-hallucination guardrails, those are good),
2) Stop the REPLACE ALL THE HUMANS agitprop and sales pitches, maybe start presenting it as a technology which will overthrow rather than empower aristocracy,
3) Stop sucking up compute silicon and driving powerful home computing to absolute unaffordability,
4) Lean really heavy into local inference and open weights, for information privacy and personal sovereignty; we [smart ones] can see that the cloud monster is the Intelligence Community's golden goose right now and we're not buying it!