AI Snake Oil: What Artificial Intelligence Can Do, What It Can't, and How to Tell the Difference

On April 17, 2025, the MIT Shaping the Future of Work Initiative and the MIT Schwarzman College of Computing welcomed Arvind Narayanan, Professor of Computer Science at Princeton University, to discuss his latest book, "AI Snake Oil: What Artificial Intelligence Can Do, What It Can’t, and How to Tell the Difference," co-authored with Sayash Kapoor.

View: https://youtu.be/C3TqcUEFR58?si=BBNvqIUQ2zannm3g


Narayanan suggests three ways to shape AI for the better:
  1. Resist overhyped and harmful applications
  2. Guardrails for specific risks
  3. Limit tech companies' power and redistribute AI benefits
He has also identified a set of established narratives about AI:
  1. Superintelligence that will usher in utopia
  2. Superintelligence that will doom us
  3. It's a fad that will pass, we should be sceptical of it,
when it's more likely to be transformative but over decades.

There's a lot more to the video but I highlighted these because even though this is a news topic, the narratives still permeate it thoroughly. Kapoor also publishes a newsletter which seems worth a rummage at least and probably merits more.

 

“Our research indicates significant complexity required for deployments of Nvidia systems in comparison to traditional data centers—cooling, configuration and orchestration challenges throughout the supply chain,” Goldberg wrote in a research report.

The analyst noted that all of Nvidia’s largest customers are also looking to design their own chips.

Interesting points for the whole industry.

Regards,
 
Scammers using AI have reinvented what used to be known in the American Civil War as 'Bounty Jumping'.

If you're a community college student in California, there's a chance that at least one of your fellow students is actually an AI bot robbing taxpayers. Recent data from the California Community Colleges Chancellor's Office suggest that these bots have stolen more than $10 million in federal financial aid and upward of $3 million in state aid between March 2023 and March 2024.

AI Bots in California Steal Over $10 Million in Federal Financial Aid
 
A replacement for silicon?

This week, the team shared what they believe is the world's most electrically conductive organic molecule. Their discovery, published in the Journal of the American Chemical Society, opens up new possibilities for constructing smaller, more powerful computing devices at the molecular scale. Even better, the molecule is composed of chemical elements found in nature—mostly carbon, sulfur, and nitrogen.

Other finds

The quantum front

Sigh
 
Last edited:
Artificial intelligence models have long struggled with hallucinations, a conveniently elegant term the industry uses to denote fabrications that large language models often serve up as fact.

And judging by the trajectory of the latest "reasoning" models, which the likes of Google and AI have designed to "think" through a problem before answering, the problem is getting worse — not better.

As the New York Times reports, as AI models become more powerful, they're also becoming more prone to hallucinating, not less. It's an inconvenient truth as users continue to flock to AI chatbots like OpenAI's ChatGPT, using it for a growing array of tasks. By having chatbots spew out dubious claims, all those people risk embarrassing themselves or worse.

Worst of all, AI companies are struggling to nail down why exactly chatbots are generating more errors than before — a struggle that highlights the head-scratching fact that even AI's creators don't quite understand how the tech actually works.
 
OpenAI, the parent of artificial intelligence service ChatGPT, has announced a new governance plan after a bitter power struggle over the business.
Boss Sam Altman said OpenAI would remain under the control of its for-profit board, while becoming what is known in the US as a public benefit corporation.
Mr Altman had put forward a similar plan in December - but without clarifying the control of the non-profit.
The update follows widespread scrutiny of the startup, which began as a non-profit and faced criticism, including from co-founder Elon Musk, that its quest for growth is pushing it to stray from its original mission of creating technology for the benefit of humanity.

 

Similar threads

Back
Top Bottom