Gary Marcus

Even criminals with very limited skills will be able to target victims at every scale due to AI

Anthropic's revenue in Q2 may not recur at the same scale next year after the collapse of tokenmaxxing, as companies like ATT turn to open source models to save costs

Big tech companies might be more screwed than expected given how toxic data centers have become

If OpenAI goes down, the damage could spread widely across the AI industry

Nvidia releasing a truly open-source Nemotron model could ultimately undercut companies like OpenAI and Anthropic

The referenced cartoon depicting AI risks or failures is likely to be highly prescient and representative of how the century unfolds

AI will not exceed the intelligence of all people by the end of 2025

Robots will not be the best surgeons on earth anytime soon

Very soon, AI systems will become robustly better than human mathematicians at solving mathematical problems

AI will struggle in writing clear, understandable mathematical proofs in the near term

AI will struggle in developing new mathematical theories in the near term

AI will not be superior to the best human mathematicians in the near term

It is uncertain whether OpenAI and Anthropic can survive the commoditization of AI

A single discovery could undermine the need for megascale data centers

Astra will very likely be vulnerable to the same problems as previous models

AGI will come someday, but pretending every new model is AGI is not hastening that moment

Astra will be a letdown for people dreaming it is imminent ASI, similar to how GPT-5 was a letdown

AI models will not deliver the same quality of results on most real-world problems as they do on math problems

There may be many corners of math where Astra is less effective

Astra-like models will likely not deliver quality results for most real-world problems despite being usable

Astra will likely disappoint like previous hyped models

Elon Musk's prediction that Optimus will be better than the best humans in surgery by end of decade is absurd and will not happen

GenAI will cause immense chaos in society

In a bust scenario, Nvidia's collateralized chips would be worth a fraction of their current value

Anthropic is unlikely to reach $100B+ revenue for the year with $85B+ remaining to achieve

Market demand for NVIDIA products may be weakening

The current state of AI insularity is not expected to end well

The AI jobocalypse is unlikely to happen anytime soon

An AI cold war with China is likely to end in an uncomfortable way for all with no clear winners

The no moat argument leading to more competitors, price wars, and scarce profits may wreck the U.S. AI industry

The business models of OpenAI and Anthropic are called into serious question by open weight Chinese models, which may kill or greatly undermine their IPOs

Massive data center investments will not pay off as price wars drop token prices to near zero, with meager profits unable to justify outlays

The no moat/more competitors/price wars dynamic will result in scarce profits and may wreck the U.S. AI industry

Anthropic and OpenAI will struggle to pull off trillion-dollar IPOs due to government controls and price sensitivity in token budgets

Alignment being solved is a distant prospect

Alignment being solved is a distant prospect

Alignment is a distant prospect to solve

When AIs can be trusted to function autonomously, it will be a big deal

When AIs can function autonomously and reliably, it will be a big deal

AGI has not arrived yet, and when it does we will stop reading about AI reliability problems

AI backlash will be significant enough to matter in the 2028 election

AI backlash was predicted two years ago and will matter in the 2028 election

AI backlash will be significant enough to matter in the 2028 election

AI backlash will significantly matter in the 2028 election

AI backlash was predictable and would matter in the 2028 election

There will be an AI backlash that will matter in the 2028 election