Graham Knapp

@simon@fedi.simonwillison.net why would Astral suddenly want to create huge amounts of busy work for maintainers and their LLMs? 😉

@nightlynx@chaos.social @bagder@mastodon.social to boldly GET where no-one has got before

@mahryekuh@hachyderm.io my AI nightmare scenario is that by mid 2027 98% of product features are adding, removing, filtering, classifying and regurgitating AI slop.

@jimbob@aus.social aiee yes AI makes it trivially easy to prepare superficially relevant applications and the AI doesn't even feel bad about blatantly lying

@_chrismay@fosstodon.org great discussion - thank you very much!

@mkennedy @brianokken just finished cleaning the floor, listening to @pythonbytes when my daughters tramped in from the muddy garden. I need a read-only house 😂

@digiglean @mathsppblog thanks for the mention - "someone I met at Djangocon" 😂

@webology reminds me of a recent bug at work with hard coded file extensions... It helps to recognise that all computer systems are fundamentally broken

@mathsppblog I'm intrigued what you are looking forward to in partial functions - is it Placeholder? @digiglean

Question for academics on Mastodon - is it easier (relatively) to recruit good students for PhDs, etc, as the job market is now so tough for graduates? In which subjects are you seeing this ? Boosts and anecdotes welcome! #AcademicChatter

@simon@fedi.simonwillison.net thank you for continuing to point this stuff out. Im sure many AI commentators would be scared of publicly criticising Anthropic or openai but it does your journalistic credibility a lot of good and I appreciate it.

@rrb@infosec.exchange ugh I had hoped that the US, being first to switch their hiring budgets to tokens, would be first to see graduates take opportunities for higher education in the interim but my new theory is that the AI jobs are hoovering up the best graduates together with all the investment funding and the weaker graduates are left doing data entry and AI training and supervision gig work

@hynek@mastodon.social awesome! Would it make sense to use this with #django or does it duplicate / clash with Django's own dependency management ?