AIs that Can’t Count the Rs in Strawberry
4–5 minutes read
News from Bloomberg
Databricks CEO Says Companies Erred With Focus on ‘Chatbots Galore’
The result … premature declarations of success….“You can train a model that’s Math Olympiad level, or can win programming contests, but it still can’t count the number of ‘r’s in the word ‘strawberry,’” he [Databricks CEO Ali Ghodsi] said. “This is the problem.”
Sept. 11. 2025
_____________________________
News from The Atlantic
The Job Market Is Hell
Young people are using ChatGPT to write their applications; HR is using AI to read them; no one is getting hired.
Corporate profits are strong, the jobless rate is 4.3 percent, and wages are climbing in turn. But payrolls have been essentially frozen for the past four months. The hiring rate has declined to its lowest point since the jobless recovery following the Great Recession.
September 8, 2025
_____________________________
News from Ars Technica
Developers joke about “coding like cavemen” as AI service suffers major outage
Anthropic outage takes down AI tools some developers rely on to create software…. One Hacker News user observed: “It’s like every other day, the moment US working hours start, AI (in my case I mostly use Anthropic, others may be better) starts dying or at least getting intermittent errors.…. [One user] recalled a joke from a previous AI outage: “Nooooo I’m going to have to use my brain again and write 100% of my code like a caveman from December 2024.”
Sept. 10, 2025
_____________________________
News from Tollbit
AI Scraping Is On The Rise. The next AI visitors won’t look like bots.
Today’s AI browsers and devtools, from Perplexity Comet to Firecrawl to Browserless, look human in site logs— showing up as Chrome while loading pages, ads, and solving CAPTCHAs….the overall increase in AI visitors puts further economic strains on sites, as their CDN costs increase to support this additional bot traffic amidst declining human referrals.
Sept. 8, 2025
_____________________________
News from Lawfare
When the Vibes Are Off: The Security Risks of AI-Generated Code
Vibe coding produces software riddled with insecurities. Will risk management and regulatory compliance, too, fall victim to the vibes?
Sept. 10, 2025
_____________________________
News frm The Rebooting
Reasons for optimism
AI is hitting a wall….the timeline for AGI – the obsession of the tech elite that will signal they’ve created their Frankenstein’s monster – is getting predictably fudged. I’ve learned to take Silicon Valley’s prophets seriously but not literally because it is a cult driven by blind belief as much as it is a hard-headed engineering culture. GPT-5 showed that humans have enough time to figure all this out. Companies are struggling to find success in implementing AI into workflows.
Sept. 9, 2025
_____________________________
News from Axios
The day is near when bad actors will use AI to hijack another AI system companies rely on
— think chatbots or agents — forcing it to go rogue, according to John Watters, a longtime cybersecurity leader and former leading executive at Google’s Mandiant. … Watters echoes a warning from another industry titan, Kevin Mandia https://www.axios.com/2025/05/13/mandiant-founder-artificial-intellience-cyberattack: The world is only months away from an untraceable cyberattack run entirely by an autonomous AI agent…. Models from OpenAI, Anthropic and other popular AI companies aren’t likely to be involved in the attack that Mandia is predicting.
September 09, 2025
_____________________________
News from RAND
Averting a Robot Catastrophe
…the authors assess the convergence of trends in robotics and frontier artificial intelligence (AI) systems, particularly the exacerbated national security risk that results from the potential for the proliferation of robotic embodiments of artificial general intelligence (AGI).
Sep 9, 2025
_____________________________
More news briefs —>
# # #
