Skip to content
Menu
Love it or loathe it, Artificial Intelligence (AI) is here to stay – and it’s getting harder to ignore. From how we generate and consume content to how industries operate and innovate, AI is reshaping the world around us. It’s making waves across sectors from media to science, transforming not just the headlines, but the way we live and work.
Each month, the Fire on the Hill ‘AI Digest’ breaks down the biggest stories in AI, exploring their broader impact on brands, industries, and everyday life.
This month we return to the top story from October and see how Atlas, the new web browser from OpenAI, is faring. Avid readers will recall the new tool allows ChatGPT to follow users across the web, understand what they are trying to do, and work to complete tasks. How was it received? Matt Honan in MIT Technology Review brands Atlas “fine”. He adds: “I asked the built-in ChatGPT to summarize an MIT Technology Review article I was reading for me. Yet instead of answering the question about the page I was on, it referred back to the page I had previously been on when I started the session. Which is to say it spit back some useless nonsense. Thanks, AI.”
While also detailing how the tool failed to fill an Amazon basket with appropriate goods and struggled to create a worthwhile Facebook post, he concludes it is “hard for me to understand why Atlas exists”. Not a great start.
Elsewhere, Imran Rahman-Jones for the BBC struggled to get Atlas to do very much of anything at all without paying for it. “What’s clear is this will be a premium product which will only work to its full capacity if you pay a subscription fee. Given we are so used to browsing the internet for free, this would require a lot of people to alter their habits quite dramatically,” he explains.
But never fear OpenAI, Comet from Perplexity has had an equally rough ride. Imad Khan reviewed it for CNET and found using it was like “a roll of the dice”. He did add a slightly more positive note: “The AI power at the core of Comet changes how I use the internet in slight but meaningful ways. AI can expand upon an article on a political candidate, giving me the ability to ask about their policies and voting record.” Ruben Circelli over at PCMag explained: “Comet’s ability to actually do stuff for you like an assistant is limited. Expect Comet to regularly get stuck or fail. And when it works, it’s usually slower than doing something yourself.” So, fair to say, AI-driven web browsers remain for the connoisseur right now, but time will tell if they grow to become a force in the future.
Industry leaders encounter headwinds for first time
How large a part might OpenAI play in that future? For the past few years, the answer has been extremely positive, but are the first cracks beginning to show? What once seemed like an unshakeable lead is showing signs of strain, according to the Economist. Growing competition, soaring costs for building next-generation models, and mounting pressure from regulators and investors are all starting to chip away at its dominance, the publication argues. Rivals are advancing quickly, offering alternatives that challenge OpenAI technologically, commercially, and ethically.
Meanwhile, the economics of massive AI models look increasingly risky. Training and deployment costs are rising fast, while returns remain uncertain. In short, what had seemed like a dominant force is now facing a pivotal moment of recalibration as the AI sector moves toward a more distributed, contested, and uncertain future.
Similarly, once viewed as nearly unassailable, questions are now being asked – if quietly – about Nvidia. The go-to supplier for the GPUs powering modern AI work, its dominance in the AI-chip market felt all but guaranteed. However, that perception has now shifted. Google recently unveiled Gemini 3, a cutting-edge AI model trained exclusively on Google’s own tensor-processing units (TPUs), potentially offering a lower-cost alternative to those who can make use of the technology. Big players — including Anthropic and potentially Meta — are reportedly considering shifting AI workloads away from Nvidia hardware.
The move chimes with the views of Michael Burry, of The Big Short fame. He argues that Nvidia, while becoming the “picks-and-shovels” supplier powering the AI ambitions of many companies, is chronically overvalued. Burry compares Nvidia to Cisco in the late 1990s dot-com bubble, a company whose stock famously once soared as the foundation for a technological revolution, only to crash when the bubble burst.
As we saw last month, he also argues that some accounting and business arrangements in the sector (like extended depreciation schedules and circular financing between tech firms and chip suppliers) inflate profits artificially and mask long-term risks.
Do you dare bet against Nvidia and the AI boom?
