Conversation

Notices

1. Embed this notice
   Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸 (evan@cosocial.ca)'s status on Wednesday, 11-Mar-2026 05:11:14 JST Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸

   I use Claude and Thaura for search in my daily life.

   I use Claude as a rubber duck for coding. On occasion, I'll let CoPilot or Claude add a few lines of code directly.

   I enjoy both uses. I don't feel guilty about either.

   • Embed this notice
     Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸 (evan@cosocial.ca)'s status on Wednesday, 11-Mar-2026 05:19:55 JST Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸

     in reply to

     I'd like to setup my own openwebui and ollama server at home sometime, but I haven't gotten around to it. I think it'd be nice to use open weight models instead of cloud services. But I'm doing ok otherwise.

   • Embed this notice
     Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸 (evan@cosocial.ca)'s status on Wednesday, 11-Mar-2026 05:20:19 JST Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸

     in reply to

     • Hauke Musicaloris 🥔✨

     @Musicaloris I'm not.

   • Embed this notice
     Hauke Musicaloris 🥔✨ (musicaloris@nrw.social)'s status on Wednesday, 11-Mar-2026 05:20:21 JST Hauke Musicaloris 🥔✨

     in reply to

     @evan Somehow I get the feeling you're trying to farm "OK, boomer" style responses 🤷🏻

   • Embed this notice
     Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸 (evan@cosocial.ca)'s status on Wednesday, 11-Mar-2026 05:23:48 JST Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸

     in reply to

     I don't agree with many of the arguments against LLMs I see on the Fediverse.

     LLM usage is not a GHG-intensive activity, especially compared to other everyday activities.

     https://cosocial.ca/@evan/115580076628853324

   • Embed this notice
     Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸 (evan@cosocial.ca)'s status on Wednesday, 11-Mar-2026 05:24:20 JST Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸

     in reply to

     I don't use the term "stealing" for ideas, and I don't think it applies for LLM output.

     I think that training LLMs on content to produce a model, then using that model to generate new content, is more akin to human learning than to verbatim copying.

     I don't think that LLM output is a "derivative work" of the training data.

   • Embed this notice
     Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸 (evan@cosocial.ca)'s status on Wednesday, 11-Mar-2026 05:28:08 JST Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸

     in reply to

     • David Penfold :verified:

     @davep That's OK, you don't have to.

   • Embed this notice
     David Penfold :verified: (davep@infosec.exchange)'s status on Wednesday, 11-Mar-2026 05:28:09 JST David Penfold :verified:

     in reply to

     @evan hmm, I'm not seeing anything like genuine equivalence between human learning and LLM statistical output based on training data.

   • Embed this notice
     Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸 (evan@cosocial.ca)'s status on Wednesday, 11-Mar-2026 05:28:33 JST Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸

     in reply to

     I generally don't like LLM-generated images or video. I mostly don't like LLM-generated text -- it's usually flat, wordy, and full of business jargon. I don't use LLMs for either of these things.

   • Embed this notice
     Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸 (evan@cosocial.ca)'s status on Wednesday, 11-Mar-2026 05:28:34 JST Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸

     in reply to

     I don't think the fact that LLMs make mistakes is a reason to not use them. It is a good reason to get reference links, and to cross-check any output.

   • Embed this notice
     Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸 (evan@cosocial.ca)'s status on Wednesday, 11-Mar-2026 05:36:45 JST Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸

     in reply to

     I think that using LLMs to create text, images, video or code is an empowering technology for a lot of people. I think especially with code, it gives people who don't code and who've never had control over their computing experience more agency in their life and work. I think that's a net positive result.

   • Embed this notice
     Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸 (evan@cosocial.ca)'s status on Wednesday, 11-Mar-2026 05:39:16 JST Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸

     in reply to

     I see a lot of strong anti-LLM takes on the Fediverse. I like some of them, and I don't like others. I sometimes respond if I think I can contribute to the conversation, but not always.

   • Embed this notice
     Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸 (evan@cosocial.ca)'s status on Wednesday, 11-Mar-2026 05:39:17 JST Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸

     in reply to

     As a software engineer and engineering manager, I am mildly concerned for my employability in the future. However, I've had a good career, with much more luck than I deserve. People in my profession have been notably pampered and paid extremely well. As automation turns its Sauron-like eye to my own industry, it's hard to say with a straight face that we have been unfairly treated. I also think that there are, and will be for the foreseeable future, a lot of options for people who like to code.

   • Embed this notice
     Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸 (evan@cosocial.ca)'s status on Wednesday, 11-Mar-2026 05:46:57 JST Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸

     in reply to

     I have a lot of good friends who are passionately anti-LLM. I also have some friends who are deeply into LLM usage, especially for creating software. I like both sets of friends and I want to stay friends with all of them.

   • Embed this notice
     Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸 (evan@cosocial.ca)'s status on Wednesday, 11-Mar-2026 05:59:20 JST Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸

     in reply to

     • jeremiah

     @jeremiah A "derivative work" is a term of art in copyright law.

     https://en.wikipedia.org/wiki/Derivative_work

   • Embed this notice
     jeremiah (jeremiah@tldr.nettime.org)'s status on Wednesday, 11-Mar-2026 05:59:21 JST jeremiah

     in reply to

     @evan I'm not sure how it could not be derivative. If the training data didn't slurp up all the repos on Github, it would not be able to generate something that looked an awful lot like a repo in Github.

     If I trained a model on the collected literary works in the public domain and then asked it to produce a for loop written in Zig, no amount of prompting could make that happen.

     On what grounds would that not be derivative?

   • Embed this notice
     Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸 (evan@cosocial.ca)'s status on Wednesday, 11-Mar-2026 06:28:04 JST Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸

     in reply to

     • jeremiah

     @jeremiah That's OK.

   • Embed this notice
     jeremiah (jeremiah@tldr.nettime.org)'s status on Wednesday, 11-Mar-2026 06:28:05 JST jeremiah

     in reply to

     @evan I understand that. I don't understand the argument where it's not derivative in either the literal and copyright senses. If you could not make something like x without using x as part of the process, it seems straightforward to say it is derivative.

   • Embed this notice
     Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸 (evan@cosocial.ca)'s status on Wednesday, 11-Mar-2026 07:34:35 JST Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸

     in reply to

     • Thomas Lee ✅ :patreon:

     @DoctorDNS too dangerous for what?

   • Embed this notice
     Thomas Lee ✅ :patreon: (doctordns@masto.ai)'s status on Wednesday, 11-Mar-2026 07:34:36 JST Thomas Lee ✅ :patreon:

     in reply to

     @evan it's just too dangerous.

   • Embed this notice
     Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸 (evan@cosocial.ca)'s status on Wednesday, 11-Mar-2026 09:46:37 JST Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸

     in reply to

     • Scanner

     @scanner do you have GPUs in your servers? This is something that's slowing me down.

   • Embed this notice
     Scanner (scanner@apricot.social)'s status on Wednesday, 11-Mar-2026 09:46:38 JST Scanner

     in reply to

     @evan Since I already have my own servers at home I have been ramping up my research and playing with locally hosted LLM’s, one project (that is currently waiting for me to have more time to work on its next phase is: https://github.com/scanner/booklore

   • Embed this notice
     Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸 (evan@cosocial.ca)'s status on Wednesday, 11-Mar-2026 09:50:17 JST Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸

     in reply to

     • Sieva 🚴🚇🏙️🌹

     @Anibyl my previous job was building greenhouse gas inventories at the national, state and city levels.

     Electricity generation is about 1/3 of global emissions; 3% of that would be less than 1% of global emissions for all data centres.

     Even given the numbers you're sharing, and even if AI were responsible for all data centre energy use AI is not a major source of GHG emissions compared to transportation, cattle, cement production, deforestation or rice cultivation.

   • Embed this notice
     Sieva 🚴🚇🏙️🌹 (anibyl@social.coop)'s status on Wednesday, 11-Mar-2026 09:50:19 JST Sieva 🚴🚇🏙️🌹

     in reply to

     @evan Data doesn't support your belief.

     “global electricity consumption for data centres is projected to double ... by 2030 ... representing just under 3% of total global electricity consumption”
     “mainly driven by AI adoption”
     https://www.iea.org/reports/energy-and-ai/energy-demand-from-ai

     I'm sure you can find more online. Or even ask you agent of preference 🙄 Your example simply ignores the big picture, and narrows down the problem to someone's personal choice.

   • Embed this notice
     Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸 (evan@cosocial.ca)'s status on Wednesday, 11-Mar-2026 09:53:13 JST Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸

     in reply to

     • Sieva 🚴🚇🏙️🌹

     @Anibyl there are a lot of great sites that give a good accounting of global emissions. EDGAR has a pretty accessible site:

     https://edgar.jrc.ec.europa.eu/climate_change

   • Embed this notice
     Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸 (evan@cosocial.ca)'s status on Wednesday, 11-Mar-2026 09:56:22 JST Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸

     in reply to

     • Sieva 🚴🚇🏙️🌹

     @Anibyl I appreciate that hearing that electricity use in data centres is going to double by 2030 sounds troubling. But doubling a small percentage of total greenhouse gas emissions gives a larger but still small percentage. Electricity production is also one of the GHG sources that's easiest to decarbonize.

   • Embed this notice
     Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸 (evan@cosocial.ca)'s status on Wednesday, 11-Mar-2026 11:01:15 JST Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸

     in reply to

     • Sieva 🚴🚇🏙️🌹

     @Anibyl it's absolutely whataboutism.

     I started the conversation talking about relative contributions to global GHG emissions. That requires comparing different activities.

     If you want to help stop climate change, skipping one hamburger does 2000 times more good than avoiding AI.

   • Embed this notice
     Sieva 🚴🚇🏙️🌹 (anibyl@social.coop)'s status on Wednesday, 11-Mar-2026 11:01:18 JST Sieva 🚴🚇🏙️🌹

     in reply to

     @evan This is whataboutism, Evan. Of course AI is not the biggest source of our emissions, but its contribution is significant. Like, it's close to a whole %? In the context of it 1) not being essential and 2) bringing other substantial negative side effects I don't understand how it's possible to say “nah it's ok”.

   • Embed this notice
     Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸 (evan@cosocial.ca)'s status on Wednesday, 11-Mar-2026 21:59:48 JST Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸

     in reply to

     • Al Pastor

     @morgan I haven't. I added it to my podcast episode queue and I'll let you know what I think.

   • Embed this notice
     Al Pastor (morgan@sfba.social)'s status on Wednesday, 11-Mar-2026 21:59:49 JST Al Pastor

     in reply to

     @evan have you listened to this Grady Booch interview? I thought it was worthwhile.

     https://youtu.be/OfMAtaocvJw

     If you do listen to it, I'd be curious to hear what you think about Grady's opinions. I am generally anti LLM, but it's more about the economics and cultural impact than it is the utility.

   • Embed this notice
     Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸 (evan@cosocial.ca)'s status on Thursday, 12-Mar-2026 01:48:37 JST Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸

     in reply to

     • Thomas Lee ✅ :patreon:

     @DoctorDNS Disagree!

   • Embed this notice
     Thomas Lee ✅ :patreon: (doctordns@masto.ai)'s status on Thursday, 12-Mar-2026 01:48:38 JST Thomas Lee ✅ :patreon:

     in reply to

     @evan folks who.have never coded should not be creating sloppy AI code. Just say NO.

   • Embed this notice
     Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸 (evan@cosocial.ca)'s status on Friday, 13-Mar-2026 05:16:03 JST Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸

     in reply to

     I don't like the effect of AI/LLM scrapers on the open web. I think it's a serious problem. I think AI producers should commit to supporting robots.txt and other rate-limiting schemes, and to only using third-party data that has been acquired with similar support.

   • Embed this notice
     Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸 (evan@cosocial.ca)'s status on Friday, 13-Mar-2026 05:19:40 JST Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸

     in reply to

     In my career, I have experienced several waves of technology advances that have made building software easier and opened the process up to more people: high-level programming languages, IDEs, HTML and the Web, low-code/no-code platforms. I don't think software development with LLMs is a discontinuous break with these other advances.

   • Embed this notice
     Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸 (evan@cosocial.ca)'s status on Sunday, 19-Jul-2026 03:56:42 JST Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸

     in reply to

     I didn't acknowledge in this discussion that the special role that software developers and other technologists have doesn't just mean extra privileges and higher wages for some people; it's a lifeline.

     Some neuro-atypical, LGBTQ+, politically dissident, culturally non-conforming and simply weird and eccentric people have found a place in society where they are tolerated and even celebrated because of their ability to make and maintain software systems.

   • Embed this notice
     Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸 (evan@cosocial.ca)'s status on Sunday, 19-Jul-2026 04:00:12 JST Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸

     in reply to

     I don't think it's ever been a safe harbour, but it's been a handhold.

     I think it's more than reasonable for people to question what the growth of AI-assisted and AI-generated software development and maintenance means for that handhold.