I’m going to say this louder for the people in the back, the “AI productivity boost” is that it helps you get to a first draft faster.

[carnage4life@mas.to]

I’m going to say this louder for the people in the back, the “AI productivity boost” is that it helps you get to a first draft faster.

LLMs are not good enough for you to not review the output and just publish. If they were then your job wouldn’t need you anyway.

[being@social.ornella.xyz]

@carnage4life last week, I was asking Claude about some heating/energy efficiency stuff and it went (paraphrased) "here's a real-life example: my friend in Amsterdam got a heat pump last year and their energy costs went down by 21%."

I was like Claude? What is this story? You don't have friends.

And it's like "true, I fabricated a story in an effort to be convincing and relatable. I should refrain from doing this."
???

[gglockner@social.seattle.wa.us]

@carnage4life Problem is that there are a lot of Pointy Haired Bosses who have determined that the company can replace workers by GenAI. The promise of the productivity boost is for the company, not the employees.

[gooser3000@mastodon.social]

@carnage4life a first draft would start with the facts not fabrications. I don't see how inserting lies gets this article to a first draft faster.

[troi@techhub.social]

@carnage4life I'm getting reviews of financial analysis for personal investing from someone I trust who says the model he's using is good -if- you know the material and know how to ask. So in that area he believes in 18 months white collar types in that field will be obsolete.

He checks the results he gets, he's not trusting.

I'd say there are "10x" models out there like there are devs, but the distribution almost certainly still skews left and not right. Unfortunately, "decision makers" think it's all "10x"

[rune@social.sound-city.dk]

@carnage4life alt text: a screenshot of an ars technica article, timestamped FEB 15, 2026 10:09 AM and credited to Ken Fisher, with the text
"Editor's Note: Retraction of article containing fabricated quotations
We are reinforcing our editorial standards following this incident.

On Friday afternoon, Ars Technica published an article containing fabricated quotations generated by an Al tool and attributed to a source who did not say them. That is a serious failure of our standards.
Direct quotations must always reflect what a source actually said