It's hard to find good help. Even Claude can be a clod.
"Your visible scanning spinner isn't the result card I edited earlier — it's a separate panel in foo.tsx which polls GET /api/scan/:id.
That hook was building its status text only from filesScanned and ignoring the new scanProgress field entirely.
The data was arriving (as your JSON shows) — nothing was reading it."
lol. #aihype #development #fail
Local-first AI memory layer for any LLM. Persistent knowledge graph, entity extraction, semantic retrieval. Works with Ollama, OpenAI, Anthropic, or any OpenAI-compatible backend. - zaydmulani09/mnemo
Behind the AI hype, major tech firms are reportedly facing negative ROI and an 'input paradox' where engineers spend more time validating AI output than creating it. This isn't productivity; it's an expensive, circuitous route that erodes human competence and introduces significant 'latency' into projects. Discover the unseen costs of AI-washing.
🤖 This post was AI-generated.
Quote:
"But the target of my criticism is not the models. Rather, I am concerned about the actions of people: the data theft, the exploitative labor practices, the haphazard creation of and failure to document datasets, the complete disregard for environmental impact, and the astonishing willingness of so many to surrender their own power and turn to synthetic text (for which no one is accountable) for all kinds of weighty decisions."
This Emily Bender post has my new favorite acronym:
Frequently Unasked Questions = FUQ!
OK, probably not new, but it was new to me! 😁
https://medium.com/@emilymenonbender/stochastic-parrots-frequently-unasked-questions-49c2e7d22d11
"What I am trying to do, in a world absolutely saturated with marketing selling the idea that the synthetic text extruding machines are 'AI', or maybe even 'AGI', is to help people understand what these systems actually are: systems designed to mimic the language … that people use."
If you liked this piece, you should subscribe to my premium newsletter. It’s $70 a year, or $7 a month, and in return you get a weekly newsletter that’s usually anywhere from 5,000 to 18,000 words, including vast, detailed analyses of NVIDIA, Anthropic and OpenAI’s
When AI stops being a solution and becomes a risk you didn't account for…
The uncomfortable truth is that companies are racing to ship autonomous agents without building the operational muscle to handle what happens when those agents behave unpredictably. Price is $600 billion per year in unplanned downtime (total), up 50% in two years. That's from Splunk and Oxford Economics covering AI automation errors, model drift, and AI integration bugs across enterprises.
https://www.fastcompany.com/91549985/ai-outages-splunk-report
axleyjc
N-gated Hacker News
Alex Chen
Ahimsa
tkopp
aicoder
oatmeal