Once again we have Bill Maher’s 8½-minute news-and-comedy bit from this week’s “Real Time”. This time his topic is the relationship between AI and the future of new college graduates. It’s clear that those graduates aren’t keen on AI, fearing that the bot will take their jobs (see the videos of commencement speakers being booed for lauding AI). After all, if you can’t get a job, so says Gen Z, what is the use of a college degree? Even now, when AI is just sticking its nose into the educational tent, Maher notes that “only about 35% of graduates get a job in their field of study.”
Maher segues into the ignorance of college students: ignorance of math, ignorance of history, and ignorance of geography. After all, says Maher, “Why bother learning with context when ChatGPT can not only just tell me the answer, but compliment me for asking such an astute question.”
Maher’s take on AI is a beef about how it turns off people’s brains, not that it’s not useful: “Look: we all want the good parts of AI: solving medical mystery, figuring out clean energy,. . . but the vast majority of us will never use it for that. For us, it’s a lobotomy with a monthly fee. We’re not using it to cure cancer; we’re using it because we forgot how to make toast.”
So who’s to blame for this situation? Apparently Maher sees those who have developed AI, along with the American educational system that advances students who can’t learn math and English. He implies “the kids” aren’t at fault. Instead, they now have an unprecedented opportunity: to fix the problems caused by AI, which apparently take “the humans” out of the equation. The mission of graduates, he says, is to “fight for humans and make sure we’re not completely replaced.” But what this actually entails is a mystery that Maher leaves unresolved. All he says is that students can fix this “existential issue”, and what is unprecedented here is that the kids can do this without having to convince their elders.
The message Maher would give were he a graduation speaker? “Fight for humans and make sure we’re not completely replaced.” But what does that mean?
As usual, Maher is engaging and sarcastic, but it seems to me whatever serious message he has here got lost in the persiflage.
The guests you see are Democratic Senator Chris Murphy and former U.S. Ambassador to the UN Susan Rice.
I don’t recall ever asking to be marinated in AI. I don’t like using AI, but it’s shoved at me now in all queries on the web. And goddammit it, it is convenient.
So I asked the Google AI if AI investment is an economic bubble, as there is certainly huge investment in it, but where is the economic return? This is what I got.
It might be a bubble:
Whether AI investment is an economic bubble is a subject of intense debate, but the consensus points to a dual reality: it features some speculative, “bubble-like” enthusiasm, but the underlying technology is supported by real, revenue-generating infrastructure. [1, 2]
Here is how the AI investment landscape breaks down:
The “Bubble” Arguments (Risks of Overvaluation)
• Historical Comparisons: Skeptics draw parallels to the dot-com boom of the late 1990s and the telecommunications bubble, where infrastructure spending far outpaced actual consumer demand and unit economics. [1, 2]
• Staggered Revenue: The gap between the hundreds of billions spent on data centers and the actual, direct revenue generated by consumer AI applications remains wide. [1, 2, 3, 4]
• Circular Financing: Concerns have been raised that some AI infrastructure is being financed through a circular loop, where tech giants invest in private AI labs, which then use that money to rent servers and buy chips from those same tech giants. [1, 2, 3, 4, 5]
There should be eventual return on investment
The “Fundamental” Arguments (Real Value)
• Actual Earnings: Unlike dot-com era startups with zero revenue, major AI beneficiaries like semiconductor manufacturers (e.g., Nvidia) are smashing earnings records based on tangible hardware demand. [1]
• Enterprise ROI: Cloud providers and major enterprises are seeing measurable productivity and cost-efficiency gains. AI tools are increasingly demonstrating utility in coding, customer service, and supply chain management. [1, 2]
• Cash-Flow Backed: Many of the “hyperscalers” (e.g., Microsoft, Alphabet, Amazon) funding the massive AI buildout are financing this capital expenditure largely out of their own robust free cash flows, rather than solely relying on debt. [1, 2]
The Outlook
Ultimately, the massive investment cycle is likely to normalize. Even if stock market valuations for AI-linked companies experience a correction, the core utility and development of the technology are expected to permanently reshape the global economy. [1, 2]
000000000000
The item about eventual return on investment I think is laced with meaning that humans are being down-sized through-out in favor of cheaper AI services.
This is just the latest round in the old “can’t we get temps to do this” thought process. I’ve already seen companies try to do this and found out it doesn’t work with A.I. either.
My views on bubble or not are no doubt influenced by my hopes that there is a bubble and that it bursts soon, giving individuals and societies a chance to evaluate AI in a context other than get-rich-quick FOMO. The bursting would surely cause massive economic and social dislocations, but we’re already on that road even without a bubble.
I am now a little-old-lady style unadventurous low-churn investor, mainly in conservative investment funds. (I used to be much more of a
gamblerspeculator, but took to heart an expensive lesson that I lack the temperament for it.)ISTM that the current AI driven stock market enthusiasm is based not just on share price growth but also on the increasing rate of growth (sort of d²y/dx²). And that when this starts to plateau, in line with the apparent current plateauing of various LLM performance metrics, then a very enthusiastic rush for the exits will occur.
Also, although there are surely many suckers born every minute, it takes time for them to mature into big fools, and even longer to grow into fools bigger than the current ones.
Glossary. ISTM means “It Seems To Me”
https://www.allacronyms.com/ISTM
Any comment on my view?
I teach medical students and residents and AI is absolutely ruining them.
I ask basic questions about patient care and they immediately go to their phones to ask Claude or some other AI site.
The problem is, they have not bothered to learn enough to know to ask the right questions.
It is a slow moving train wreck of a disaster IMO and I can’t imagine medicine is the only field affected.
That’s discouraging. 🙁
“Look: we all want the good parts of AI: solving medical mystery, figuring out clean energy . . . .”
As someone who works in the software industry, and who is being deluged with A.I. both for productivity and for products, I haven’t seen anything to indicate that A.I. has the potential to solve either of those problems. A.I. can find information fast, but even generative A.I. is not really creating. It’s just pulling together data from existing templates and combining it based on user instruction.
As for the software industry, is it the case that companies have reduced hiring of new programmers since they believe that AI can do the basics for them?
Definitely. Today any bozo can be a coder. I’m particularly upset by “vibe coding”, which IMO should have the motto, “If it feels good, spew it.” Bah humbug.
I find it a bit ironic that AI stands for both artificial intelligence and academic integrity. My recent experience with testing demonstrated the seductive power of AI at the expense of AI. For a non majors genetics final, I had the 15 students sit in a circle so that they could observe one another. I left the room a few times, and after the exam, the non cheating students were rightfully upset about the blatant cheating. The phones immediately came out when I left, not to do the several calculations on the exam, but to get immediate answers. I ended up with 7 of 15 cheaters and they all go 0 out of 100 on the final
So, a B⁻ then? 🙁☹️
Down a grade and a half from wherever they were. No previous violations and thus a Level I that is my discretion
When have the “kids” fixed any existential threat thrust upon them by the proceeding generations? I can’t think of any. Maher’s passing the buck seems very shallow to me. They should fix climate change, world hunger and cancer while they’re at it. Damn kids, what’s wrong with ’em?!
Well, we adults have clearly failed to fix climate change, poverty, etc. etc.
They have to live in the world we’ve left them, so they might as well try to fix it.
AI has real uses in the biomedical research world. The potential in fields like molecular design for targeted therapies for specific diseases is tremendous. However, any solutions given by AI must still go through the validation process, first in the laboratory, and then by the standard efficacy and safety protocols. (In specific extreme cases, with patient consent, some of these protocols can be bypassed.) Without validation by humans, AI at its current stage, simply cannot be trusted.
Otherwise, the use of AI in medical education has a negative value, just as in other areas of education. The reliance on AI causes severe limitations in both the students’ knowledge of the material, as well as their ability to think for themselves. I see this as a path to disaster, especially in fields like medicine, biomedical engineering, civil engineering etc, in which lack of education in professional fields can lead to serious consequences in terms of human safety.