It's not psychosis at all. The CEOs likely know more than the average slashdotter. What might look like psychosis is much more reasonably explained by the chase for stock bonuses. AI-heavy companies like Nvidia and OpenAI would lose most of their value if AI popped, and their CEOs would lose many billions of dollars. They need to promote a rosy AI future because their they would drop from insanely rich to merely extremely rich otherwise. Even companies like Google, Apple, Meta, Amazon, and Microsoft feel the need to promote AI because they envy the stock returns of the pure-AI companies.

What do the CEOs know? That some AI domains are already extremely successful (e.g., factory monitoring and maintenance), selectively successful (e.g., financial analysis, ads), or soon to be successful and life-altering (e.g., medical diagnosis, medical treatments). Some domains are mostly not currently successful and have murky outlooks (e.g., software coding). Some domains are intended to enable future domains (e.g., frontier models) or pursued as halo projects. The people who hail AI as the precursor to the singularity and the people who deny all current AI successes are both blind.

Exactly. This is a very good sign. Despite overt and implicit opposition from the current US government, wind and solar continue to grow. That suggests strong inherent economic motivation.

"Siri should finally be able to understand more personal context, have on screen awareness, and be able to take action in apps for you. This'll finally be made possible thanks to Apple's new partnership with Google, where Apple will be using Gemini-diffused models hosted on Private Cloud Compute to power Siri"

So, basically Apple is marketing existing tools as Apple and Siri? The good thing about selling AI is that most consumers don't understand what's happening underneath the chatbot interface, so Apple can claim to be making AI innovations, and some people will believe it.

If the entry bar is a string of literals representing keywords to search for, then the response should be the old list of links to associated pages. However, if the entry bar is a conversational request, then treating "disregard" as a request rather than a literal search term is reasonable. Of course, it might be argued that no one would actually bother to approach Google to tell them to disregard a request. However, that's perhaps a matter of conversational protocol. What would a human do if someone approached and said, "Disregard"? There are likely other ambiguous or awkward instances of conversational protocol that would be misinterpreted.

The main problem is that the entry bar is intended to field both search requests and conversational requests at the same time, but there is currently no syntax or protocol to distinguish the two. Simply relying on context is insufficient because humans would have the same parsing challenges. Maybe putting quotes around literal search terms is one way, but quotes already have special meaning in search terms.

We can blame social media, billionaires, idiots, etc. However, one big thing that needs to be called out more is an increasingly authoritarian US government that is on an audacious and so-far unchecked binge of canceling. This Friday, Stephen Colbert is being canceled because Trump wanted it.

AI already works, but only in some areas like factory management, financials, medical, and ads. Those areas have already seen implementation, added functionality, and profits. However, Meta doesn't sell into those areas, aside from ads. Meta seems to be flinging people around trying to find something that sticks. However, Meta's AI incompetence doesn't erase the already existing AI successes in other companies.

Is this true of MS programs that don't require writing a thesis or research or publications? Many MS degrees, even from name universities, are purely taking more courses.

If this harsh reality is indeed the reality, then this dystopian nightmare is guaranteed to be temporary because eventually the tech will be shown to not work. So, even though people will suffer in the meanwhile, the problem will take care of itself over time. The real fear is not that the AI doesn't work but rather that the AI does work to at least some extent.

Earnest question: Is this laptop snooping to train AI replacements a Meta thing, or is it happening at other companies?

It also begs the question of the utility of grades. There's the matter of college awards, but those are largely participation trophies. The more important matter is getting a job. Do higher GPAs and maybe the awards associated with those GPAs lead to better jobs? You'd think that if employers knew about the grade inflation that they would be smart enough to discount the significance of GPAs.

Even absent any grade inflation, should employers consider GPAs as something that is predictive of workplace success? Although some hiring managers and HR screeners do consider GPAs, I've always discounted them when interviewing candidates. The other items on the resume (like work experience, projects, papers, etc.) are far more significant to me.

Maybe the real reason Harvard is considering limiting grade inflation is the hit to Harvard's reputation. Whether it's Harvard giving out A's like candy or Stanford giving out disability accommodations like candy, it's embarrassing to a university that thrives on universal admiration. This is the real reason, to keep up the external image.

It's interesting that although the article avoids stating causation, it does use the word "help," which is also an extremely strong statement when the data only shows correlation. Social-economic level, especially relating to income and assets should be the very first obvious factor analyzed. There's a reason that rich people support and participate in the arts more than poor people. Rich people have the spare time, spare money, and often less stress (economic, health, etc.). For example, some high schools require a lot of money (up to $5000 for one school in our area) per season to participating in marching band, and that doesn't even include buying the expensive instrument. Poor people have to work more hours for lower pay. When they do have spare time, simply destressing is more important than indulgence in the arts.

Insider trading without all the large institutional investors or otherwise rich people. It's those large or powerful investors that mostly keep the insider trading in check. Otherwise, the stock market would be just like Polymarket and Kalshi.

It's 65k per year plus another 20k with graduate degrees. So 85k per year. Over 10 years, that's 850k. About 2/3rds of the H1-B visas are computer and IT jobs, so, say about 550k. That's out of the roughly 6.5 million total computer and IT jobs. Or about 8.5% (if all the H1-B computer and IT workers stayed in these types of jobs). That's definitely enough to affect wages and job availability.

Nvidia is lagging its competitors with respect to stock price appreciation. However, it's still outperforming its competitors in revenue, margin, and future demand. While Nvidia's growth has indeed slowed down, it still is growing faster than Intel or AMD, either looking at percentage or dollar growth. Intel's sales are flat. While AMD's sales are growing by about 40% over the last year, Nvidia's sales have grown by 65%, despite being much larger to start with. AMD's PE ratio is now an astounding 151! Intel's PE is undefined since its profit is negative.

How about for questionable slop? Most of the AI stuff is perhaps detectable by highly trained individuals, but most readers are not highly trained. Also, many readers eagerly slurp up human-created slop of obvious low quality and accuracy, so why would there be a rejection of AI-created slop of obvious low quality and accuracy?

AI slop already works for revenue generation. That's why it's gaining steam. I'm not advocating for AI slop, but it's obvious that it's working, maybe not for the readership public but definitely for the generators.