GPT 4.5 is no GPT-5

Investors should be worried

“Draw a wine glass overflowing… fill the glass to the top … increase the level of the wine to the top of the glass”, elicited over the course of 3 prompts from Gemini, by a neuroscience friend

GPT-5 was supposed to blow everyone’s minds, solve hallucinations, fix reasoning, and maybe even take over the world. But none of that has happened; instead, efforts to build it were delayed, exactly as I predicted a year ago.

But here is the big news. At any moment, perhaps today even, OpenAI will finally deliver! [Checks notes]. Actually, at any moment, we will finally get GPT 4.5, which most certainly wouldn’t live up to expectations around GPT-5 — hence its rather tentative name.

Still, fantasies are as ever running hot. On X, one hypester wrote “gpt 4.5 is a[n] agent. gpt 4.5 has human level vision.” Another hypester wrote that “the acceleration curve is f___ing vertical now. nobody's talking about how we just compressed 200 years of scientific progress into six months. every lab hitting capability jumps that would've been sci-fi last quarter. we're beyond mere benchmarks and into territory where intelligence is creating entirely new forms of intelligence.”

Don’t count on it.

GPT 4.5 is unlikely to revolutionize AI

In reality GPT 4.5 is unlikely to meet high expectations, per The Information:

Mixed results are, of course, what I have been warning about all along. In “Deep learning is hitting a wall”, I warned that problems with reasoning and hallucination wouldn’t go away simply as a function of scaling, and they haven’t. In my predictions about o3, I predicted that new enhancements would work in some domains and not others.

All this looks to be true for GPT 4.5 as well. Certainly it seems to be true of Gemini:

Astonishing error from Google, recently reported, now patched. But we can sure expect more errors of blatant discomprehension to emerge.

The latest Claude and Grok releases aren’t exactly setting the world on fire, either.

The truth is that scaling just isn’t producing the fruit that it once did. This just in, from Fortune:

It’s not just OpenAI either; it’s the whole field, as I discussed in my own recent Fortune article:

And meanwhile GenAI has had very little impact on GDP, despite initial enthusiasm. Even Satya Nadella sees that now.

Will Nvidia stock continue to defy gravity?

All this should matter a lot for Nvidia, which is announcing quarterly earnings shortly. There might even be a bump today.

But IMHO people investing their money in Nvidia shouldn’t be looking just at quarterly reports (which might be good, since people continue to buy their gold rush shovels in bulk—for now).

Investors should also be looking at the prospects for “solving” AI with pure scaling, which is after all much of the vision Nvidia is selling.

And for that matter people should be comparing GenAI revenue (on the order of low tens of billions per year, pre price wars) with GenAI expenditures (order half a trillion dollars to date), and doing the math, especially given the price wars induced by DeepSeek. Is the demand sustainable, given that math?

Right now, the math for profitability doesn’t work, influence on productivity of GenAI has been minimal, and prospects for pure scaling aren’t looking great.

Nvidia has gotten rich despite all these concerns, but the basic problems of AI remain unsolved.

Gary Marcus apologizes for repeating himself with minor variations. But minor variations seems to be all that is emerging, despite the enormous investments.