AI isn't panning out

 From an article in Futurism, Microsoft's (to give the latest example) sales pitch isn't panning out:

Regardless, the dustup suggests that enterprise customers are far from convinced that large AI agents are ready to autonomously complete complex multistep tasks. It’s yet another indication that companies are struggling to convert the enormous hype surrounding generative AI into actual revenue, a concerning trend considering the billions of dollars AI companies are burning through right now with no end — or return on investment — in sight.

This is, of course, because it continues to absolutely not live up to the hype: 

Researchers at Carnegie Mellon University found earlier this year that even the best-performing AI agent, which was Google’s Gemini 2.5 Pro at the time, failed to complete real-world office tasks 70 percent of the time

What I find incredibly alarming is how many people who are in positions of power and authority refuse to look critical at all on this matter. We're continuing to see it pushed across the education sphere, including (particularly of concern to me) in matters of school finance. 

Consider what a 70% failure rate looks like for school budgeting. 

Office of the Inspector General on FY23 in Brockton

 I am just seeing this, but for those who read or follow such things: the Office of the Inspector General has issued a report on Brockton in FY23. You can find it here.
Coverage in MassLive, Boston Globe, and The Enterprise

Two ways the news isn't good today

 Happy Monday back from Thanksgiving! 
Two pieces of not great news today:

• We are just failing all over the place on vaccinations, and that has consequences. Last week, the United Nations warned that 30 million children around the world have missed measles vaccinations, and "...outbreaks continue to rise sharply." To wit:
  

  In 2024, 59 countries experienced large or disruptive outbreaks – almost three times as many as in 2021 – and a quarter of them had previously eliminated measles.

  At the same time, here in the U.S., whooping cough numbers are outpacing those of last year. A third baby in Kentucky died last week. 
  
  

• I will admit the phrasing here made me laugh, but it turns out an "old fashioned web search" is better for learning than using large language models (generative AI). This is from research published in October through the National Academy of Sciences. 
  And at the same time, the argument schools have been hearing on why students need to be 'taught' AI seems to not be panning out: AI use in businesses has flatlined.
  

  Three years into the generative-AI wave, demand for the technology looks surprisingly flimsy.

  That one's from The Economist, by the way, the "we're looking to see if there is money to be made" coverage, so perhaps we can take that to the bank?