Generative Artificial Intelligence (AI) tools are awesome and I use them every day. They are massive boost to coding and they make learning new things much easier. However, there are a lot of claims about AI that seem to be far ahead of where the technology is currently at. The claims that Artificial General Intelligence (AGI) is a matter of months away and that its going to take our jobs sound credible and is not meeting much resistance in the main stream media. Here are some viewpoints from prominent commentators offering more balanced counterpoints to the extreme positivity of the tech "hype bros."
AGI Is Not Coming Soon
Gary Marcus writes on substack and has been warning us for sometime that Large Language Models (LLMs) alone are not going to get us to AGI. With a PHD from Massachusetts Institute of Technology Gary has been working as an academic and journalist in the field for a long time and his views are highly respected. This post articulates many of his key points in relation to LLMs:
In this YouTube video he is interviewed on Steve Eisman's podcast discussing the main talking points:
It's easy to dismiss Gary as a sceptic and a cynic, but I don't see him that way. He does believe that we can get to artificial general intelligence, but just doesn't believe that LLMs alone will get us there and regularly cites Nuerosymbolic AI as playing an important role in the future. This current push to AGI through LLMs alone does seem short sighted and is unlikely to get us all the way to AGI.
They're Lying to You
Carl Brown comes across as an experienced, well read person who is simply paying attention to what's being said and doing his own analysis. His years of experience lend credible, reassuring reflections on the cycles of technology over the last thirty years and provide a more realistic counter point to the irrational sounding claims and predications of leading Generative AI companies.
Carl's YouTube channel, The Internet of Bugs, rose to prominence after debunking some of the early claims made by the company Cognition when trying to sell Devin the advertised first AI software engineer. Apparently Devin is capable of starting it's own side hustle by solving problems advertised on Upwork:
- Devin's Upwork Side Hustle - Original Video
- Debunking Devin - Internet of Bugs Analysis
- Devin didn't solve my computer vision project - Original Upwork Poster Response
It's worth going back through the videos just to realise the degree to which companies like Cognition are claiming things that simply aren't true. In the YouTube video below Carl is explains how the term AI is not clearly defined and the meaning is being abused to pull the wool over our eyes:
Carl does appear to be impressed with the technology, AI tools are impressive and do add value, but he seems very frustrated with the degree to which companies are wilfully lying to us in order to try sell us something.
The Numbers Don't Add Up
Ed Zitron is a bit of a legend and his articulate, passionate diatribes are entertaining and informative. He is another one doing the research and digging into the small print of the claims being made by big tech companies leaders. Ed writes a newsletter, Where's Your Ed At, and also has a podcast, Better Offline, where he shares his research into the headline stories that are making waves.
Ed adds more support to the above ideas but its his deep research of the company financials that really stands out for me. The massive valuations for companies which aren't making a profit and have very dubious financials with circular funding and capital requirements just to run the service are really concerning but as long as "number go up" people are closing a blind eye to some serious underlying issues. This post, which shares this articles title, is a good place to start:
In this video, Carl Newport and Ed review all the AI stories of 2025 month by month:
Ed doesn't pull his punches and he can be quite scathing when he talks about leading AI figures like Geoffrey Hinton and Yann LeCan, but his research is meticulous and quite damning and well worth paying attention to.
Summary
I regularly use the Generative AI coding tools and I find them very impressive but the claims of leading AI proponents seem a bit far fetched. I've found the commentators mentioned in this article interesting, informative and entertaining and they provide a more plausible counterpoint. There's a lot of myths being peddled by people with a vested interest in AGI being almost here and the mainstream media seems to be further fuelling these claims rather than challenging them. Thank-fully some professional academics, developers and journalists are paying attention and trying to correct the crazy. Integrating AI into the business world is going to be an interesting challenge as people work out what it can and can't do and I fully expect carnage as people re-organise around something that is right 80% of the time. It's sure to be a brave new world!
back