Genuine Scepticism on AI Hype

From Abhivardhan, our Chairperson

Okay, I came across this post on X/Twitter of

Gary Marcus

(do follow

Marcus on AI

) and I am visibly shocked but not that surprised.

Context: Gary is known to make reasonable assessments and commentaries on the state of Generative AI and why OpenAI's all "products" have been in preview mode for quite long, and why GPT-o-1 or GPT-5 "potential" is misleading simply because they are unprepared, half-baked products.

Now, it seems that many US-based retail investors or consumers (maybe) are doing trading around the chances that Gary would allegedly "dunk" and say the potential GPT-5 fails in the most miserable ways.

Also, if you wish to know why it is overrated to claim that GPT-Orion-1 by OpenAI is perfect and workable (it was launched merely few days ago, not even a week before), check out the second screenshot above in my post and read this amazing newsletter by

Michael Spencer

:

OpenAI Keeps Releasing Prototypes & Previews of Actual Products

My larger point is that you do not need to be delusional about GenAI models and yet support genuine AI policies and genuine AI innovation by being a healthy sceptic.

Honestly, the AI ecosystem is so nascent in India (#Bharat), I would not expect that kind of hype around these issues. But I definitely see a wave where people are still supporting AI hype and not investing in genuine AI solutions.

Anyways, for the US having a spirit of hedging on free market bets on anything, I'd hope Indians to act more responsible.

To conclude, I am happy to share a LinkedIn comment from the post featuring the same content by

Michael Spencer

from

AI Supremacy

:

“Just like in investing, you need to listen to optimists, pessimists and the realistis. Truth is often at the intersection of all of these viewpoints. Personally I find Marcus sometimes over the top, just like I find Mollick's AI evangelism over the top. Those voices keep repeating their same old because that's what their personal brand is tied to at this point.

So algorithms are really impacting how even people talk about AI, and this is occuring at scale. Even the papers where we see some of the data is partially faked. So there are levels of deception even as the bulk of the AI investments goes to just a few concentrated names. It's not good for the world.”

More ISAIL Updates