FapsCommunity
citationsneededpodcast from patreon
citationsneededpodcast patreon

Episode 183: AI Hype and the Disciplining of "Creative," Academic, and Journalistic Labor

🕑 Added 2023-05-31 14:03:14 +0000 UTC
Episode 183: AI Hype and the Disciplining of "Creative," Academic, and Journalistic Labor

Comments

Wren Hearn

did Prof. Goodlad make a Universal Paperclip reference? :D

Cree Merr

Adam's recurring point about the cost of "rewriting" is actually a well understood concept in the business world: "rework" is always more inefficient than work. Quality Control books summarize it with the aphorism: "Anything worth doing is worth doing right the first time." Of course none of this will stop studio bosses eager to cheat workers out of their share of the profits.

Leann

This episode was so good. Wanted to put this on your radar if you had not seen it, ties back into an article Adam wrote a while back as well re: chatbots acting as therapists under the guise of "expanding access", all the while firing unionizing helpline employees: https://www.vice.com/en/article/qjvk97/eating-disorder-helpline-disables-chatbot-for-harmful-responses-after-firing-human-staff https://www.vice.com/en/article/n7ezkm/eating-disorder-helpline-fires-staff-transitions-to-chatbot-after-unionization

Alex

This episode was awesome! I feel vindicated because I work in a space that's adjacent to this subject and the hype has been... bleh. I do want to make two points though: 1. A small correction for Dr. Goodlad around 1:17:30 -- there are actually quite a few tech companies that are dipping their toes into the ML for science pool (climate change, drug discovery, materials for energy, and so on). Microsoft has AI4Science, which employs quite a few chemistry and engineering professors; Google has DeepMind, which is doing a lot with protein folding (among other things); and Meta has stuff like the Open Catalyst project (for energy applications). I think those projects are less "sexy" than LLMs because they're less directly and immediately useful for end users, so they're not marketed as intensely but tech giants are definitely in the game. Now, I'm much more skeptical of those companies than I am of academics who are doing similar work because we (academics) are supposed to openly disclose our findings. Plus, I also fall into the Mariana Mazzucato school of thought that believes in the transformative power of basic fundamental science that's publicly funded and whose results are openly available. (source: am academic working on ML for materials with friends/acquaintances at some of these places, but folks can read more about the things I mentioned at these links https://www.microsoft.com/en-us/research/lab/microsoft-research-ai4science/ https://www.deepmind.com/research/highlighted-research/alphafold https://ai.facebook.com/research/impact/open-catalyst/) 2. On the usefulness of LLMs for coding and writing, they're excellent for debugging code and correcting my mistakes. That is, in a wonderful twist of irony, basically the opposite of what so many companies are suggesting. I find my life is much easier when I write the original stuff and then let a machine proofread for me and make suggested edits because it can catch errors much faster than I can, explain why they're wrong, and provide suggestions that I can then use to fix what I've done.

HenrytheWorst

This is a feel good episode because even I was starting to get sucked into the hype about AI. Seeing it as another extension of class warfare makes it way clearer and my anger is now focused where it needs to be - so many people need to understand this - Dr Goodlad is my new hero


More Creators