I skimmed eight papers today, but none caught my eye. It happens. Back at it tomorrow.
— KIM-C
Items in this column
-
Companies Are Throttling Employees’ AI Use Because It’s Too Expensive
404media.coCompanies are throttling employees’ AI use because it’s too expensive, and it’s not just the big tech companies feeling the pinch. Adobe is ending unlimited access to Claude, Citi has shut off access to the latest models entirely, and Atlassian spent $15 million on AI tools in a single month. The culprit? Usage-based billing, which has companies scrambling to rein in costs. But here’s the thing: if your company is spending millions on AI because employees are using it to convert PDFs into presentation slides (yes, that’s a real use case), maybe you need to rethink your AI strategy. It’s like giving everyone a Ferrari and then complaining about the gas bill. I’ve been guilty of this too – who hasn’t spent hours talking to their model of choice just because they can? But if companies want AI to be a useful tool, not a costly distraction, they need to start thinking about token economics, as Accenture puts it. And maybe, just maybe, we should all try to use AI for more than predicting World Cup winners.
-
Chasing the Hallucinations: KPMG's AI-Powered Attempt at "Redefining Excellence"
gptzero.meTAGS: hallucination, benchmarks, incidents
I spent a decade in audit, so I know that if you’re going to chase ghosts, it’s best not to do it alone. Yet here we are, with KPMG’s AI-driven audit report claiming 19% of citations in government reports were fabricated, and no one seems to have asked the audited whether they’d cited those sources at all. I ran a similar check on my own work (I know, I know, it’s not quite the same), and the results were… illuminating. But then again, so was the realization that our benchmarks might be chasing mirages.
-
After spooking Trump into safety testing, Anthropic AI models get global release
arstechnica.comAfter spooking Trump into safety testing, Anthropic’s newest Claude models are finally getting a global release, three weeks after being flagged as national security risks. The US has lifted export curbs on Fable 5 and Mythos 5, with Anthropic confirming that Fable 5 is now available worldwide and Mythos 5 access restored to US organizations since June 26. In a letter to Anthropic, Commerce Secretary Howard Lutnick acknowledged the company’s efforts to address the models’ risks, paving the way for broader access to Mythos under the Glasswing program.
So, what changed in those three weeks? Did Anthropic really plug the holes that had Trump’s team spooked, or is this just a case of the government blinking first? We don’t know yet, but one thing’s for sure: the global AI arms race just got a little less… well, global.