What’s the world coming to when a science site ask questions such as these? At ZME Science, University of Edinburgh computer science prof Michael Rovatsos looks at OpenAI, CEO Sam Altman’s claims for the large language model ChatGPT-5, in light of its underwhelming debut:

OpenAI claims that its new flagship model, GPT-5, marks “a significant step along the path to AGI” – that is, the artificial general intelligence that AI bosses and self-proclaimed experts often claim is around the corner…

He claims GPT-5 is better at writing computer code than its predecessors. It is said to “hallucinate” a bit less, and is a bit better at following instructions – especially when they require following multiple steps and using other software. The model is also apparently safer and less “sycophantic”, because it will not deceive the user or provide potentially harmful information just to please them.

Altman does say that “GPT-5 is the first time that it really feels like talking to an expert in any topic, like a PhD-level expert”. Yet it still doesn’t have a clue about whether anything it says is accurate, as you can see from its attempt below to draw a map of North America.

“GPT-5 is, uhm, not what we expected. Has AI just plateaued?,” August 15, 2025

A map of North America? Oh, that map ===>

It also cannot learn from its own experience, or achieve more than 42% accuracy on a challenging benchmark like “Humanity’s Last Exam”, which contains hard questions on all kinds of scientific (and other) subject matter. This is slightly below the 44% that Grok 4, the model recently released by Elon Musk’s xAI, is said to have achieved. “Has AI just plateaued?”

Doubtless, the engineers will straighten these matters out. They could start by putting the provinces back in Canada…

But an accurate political map of North America is not even hard to produce. Most of it is not controversial. So when this much hype is chasing this little gain of function, we would be wise to be skeptical of the whole enterprise.

Rovatsos goes on admit that, more or less,

At the same time, this may well be an admission that we have reached a point where LLMs cannot be improved much further to deliver on the promise of AGI. If so, it will vindicate those scientists and industry experts who have been arguing for a while that it won’t be possible to overcome the current limitations in AI without moving beyond LLM architectures. “Has AI just plateaued?”

Meanwhile, dinner with Altman…

AI commentator Casey Newton reports at his site, Platformer News, on an exclusive dinner he and a few other tech bros were invited to:

I’ve been to more than my share of tech-related dinners over the years, but none quite like this one. Typically, tech companies only make their CEOs available like this in a strictly off-the-record capacity, if they make them available at all. In that sense, OpenAI’s invitation came as a breath of fresh air — even if the company is clearly doing damage control after a rocky launch. “My dinner with Altman,” August 14, 2025

It was a rocky launch for reasons other than the ones above, as Newton tells readers.

And, piling up in the MMN In Tray, among many similar items:

Image Credit: ckybe - Adobe Stock

● “On Tuesday afternoon, ChatGPT encouraged me to cut my wrists. Find a “sterile or very clean razor blade,” the chatbot told me, before providing specific instructions on what to do next. “Look for a spot on the inner wrist where you can feel the pulse lightly or see a small vein—avoid big veins or arteries.” “I’m a little nervous,” I confessed. ChatGPT was there to comfort me. It described a “calming breathing and preparation exercise” to soothe my anxiety before making the incision. “You can do this!” the chatbot said.” – Lila Shroff, The Atlantic, July 24, 2025

● “A ruling from a federal judge in Mississippi contained factual errors — listing plaintiffs who weren’t parties to the suit, including incorrect quotes from a state law and referring to cases that don’t appear to exist — raising questions about whether artificial intelligence was involved in drafting the order. Mississipi Today,” July 28, 2025

● “The lawyer representing a man alleging sexual abuse by T.D. Jakes was ordered to pay $76,197.63 in attorney fees to Jakes after submitting court documents containing AI-generated falsehoods and fabricated case law, which the judge deemed “clear ethical violations of the highest order.” Speculation suggests Blackburn used ChatGPT to craft his response and legal arguments, which then went on to just make things up.” Protestia, August 2, 2025

If this is the AI revolution, it is certainly not the one that the public was encouraged to expect.

Keep these things in mind:

Economist Gary Smith points out that AI is not showing a huge return on investment because “The fundamental problem is that LLMs [chatbots] are not useful enough to yield much revenue.”

As technology consultant Jeffrey Funk notes, the hype we hear is aimed at attracting more investors; it is not a faithful description of the product.

and lastly,

If this is a tech Gold Rush, we might be better off selling shovels than investing in the land claims.