@toldtheworld

13 Followers
37 Following
207 Posts

..to invest in getting the underperformers to perform, because the psychological safety of the whole team mattered and there is a risk when hiring someone new anyway (and hey, they're not happy with the previous hires, so what says they won't hire poorly again?), and you're already going to spend weeks or months looking for someone and onboarding them, so why not work with what you have? The hirer looked surprised.

We mutually decided not to continue. I want a world where people are valued.

I recently interviewed for an engineering manager position at a startup and one of the first questions was how much I wanted to fire underperformers. I explained that firing meant management had a problem, either by not screening correctly or by not giving engineers the support and environment they need to perform, and that I could do it once but would always consider it a management problem, not an individual contributor's, so they could not expect me to do it regularly, preferring instead..1/2

https://tuentality.wordpress.com/2026/05/15/which-gen-ai-provider-should-you-use-to-minimize-your-carbon-footprint/

"Claude and Meta appear to be the least polluting[1]. A true believer, “native”, use-loops-and-subagents-all-the-time user of ChatGPT and Codex is responsible for burning the equivalent of at least 32 liters (7 gallons) of car fuel every month[2] — about 30 times more than a casual user, who would burn through only one liter (using Claude or LLama on AWS for the same tasks would cut emissions by half). The energy grid in China is, apparently, still quite dirty [...] #ai

https://tuentality.wordpress.com/2026/05/15/which-gen-ai-provider-should-you-use-to-minimize-your-carbon-footprint/

"Claude and Meta appear to be the least polluting[1]. A true believer, “native”, use-loops-and-subagents-all-the-time user of ChatGPT and Codex is responsible for burning the equivalent of at least 32 liters (7 gallons) of car fuel every month[2] — about 30 times more than a casual user, who would burn through only one liter (using Claude or LLama on AWS for the same tasks would cut emissions by half). The energy grid in China is, apparently, still quite dirty [...] #ai

I'm working on a book of deeply computer-related poems, each paired with a vignette. I think they fit perfectly here in Mastodon!

This time around: Buffer overflow. Your wife is sleeping with another man!

#writing #writingcommunity #poetry #poetrycommunity #computerscience #smallstory #fiction #shortstory

I used to ensure my job applications and CVs had no typos or grammar errors, but now they are signals that my cover letters are not mass produced ¯\_(ツ)_/¯
Does anyone here have any idea how to go about calculating the approximate environmental cost per token use of different #LLM models? (I imagine energy consumption would be the beginning, but where to get that information for infrastructure providers like AWS Bedrock?) #environment #co2 #accountability #aws #azure #gcp
I think I found a good use case for an #LLM: I had an ancient export of recipes from #GourmetRecipeManager in HTML that I wanted to make more accessible using #Cooklang. I created an AI skill with cooklang specs (and a command to validate the output with cooklang #cli) and used it with Opencode to automate the conversion of the recipes. The results are pretty good (#MiniMax M2.5 Free), especially since the source material wasn't perfect. LLM output is acceptable when exactitude is not required.

I'm worried about AI psychosis. Specifically, I'm worried about the psychosis that makes "capital allocators" spend *$1.4T* on the money-losingest technology in human history, in pursuit of a bizarre fantasy that if we teach the word-guessing program enough words, it will take all the jobs.

--

If you'd like an essay-formatted version of this thread to read or share, here's a link to it on pluralistic.net, my surveillance-free, ad-free, tracker-free blog:

https://pluralistic.net/2026/04/13/always-great/#our-nhs

1/

I've noticed many people now write like Claude, and I think part of it is unconscious mimicry. A classic pattern: "no this, no that. Just meh." I've been reviewing a lot of generated text and even I am starting to adopt the writing style, which bothers me for the same reason eating pasta every day does.