@peterlo.bsky.social
PR at AWS. Kpop DJ & podcaster. Talk to me about responsible AI, semiconductors, or your favorite cyberpunk novel.
I'm halfway through this article & like (and agree with) the points made thus far.
But I want to have an aside on how people uncritically use the term "foundation models" & "reasoning," when it is very likely that the models memorized all these benchmarks. π§΅
prospect.org/power/2025-0...
OpenAI exec saying "ChatGPT image generation enables so much creativity!" quoting a post about internet users using ChatGPT to turn popular memes into poor imitations of Studio Ghibli animation.
Honestly, I can think of few things *less* creative than doing derivative versions of popular memes because everyone else is doing the same thing. There's no joke or cleverness to the mashups even. Just imitations of imitations in an imitative style because that's what everyone is doing.
27.03.2025 15:22 β π 1272 π 138 π¬ 9 π 46
Some AWS work humor.
21.03.2025 19:56 β π 0 π 0 π¬ 0 π 0
βOpen source models are essential for the U.S. to win the AI race against China and ensure American AI dominance,β Meta writes in its submission, which the company provided to me. βExport controls on open source models would take the U.S. out of that race, allowing Chinese companies like DeepSeek to set the AI standard on which the world builds and embedding authoritarian values in the global tech ecosystem.β Meta does not mention that researchers have been caught using its open-weights Llama model to build applications for the Chinese military. But perhaps the democratic values embedded in Llama will serve as a bulwark against whatever China plans to do with it.
Trump's call for comments on an 'AI action plan' resulted in the big tech companies asking for favors they've wanted forever β and they're using the emergence of DeepSeek to create a possibly dangerous sense of urgency www.platformer.news/ai-action-pl...
19.03.2025 00:35 β π 1076 π 277 π¬ 20 π 25I'm feeling a little judged for enjoying my raspberry nutella gochujang miso biroche, but there's a few spots in financial district. There's a boichik near market & Battery, and then your choice of either philz or an indie coffee shop on the left & right. Bluebottle has midwife and the baker, too.
18.02.2025 17:47 β π 0 π 0 π¬ 0 π 0
This new BBC research shows some of the problems that AI chatbots have with summarising current affairs www.bbc.co.uk/news/article...
12.02.2025 11:36 β π 5 π 1 π¬ 0 π 0
Google auto opted in *every* paying Google Workspace user to Gemini. Also adding a nice increase in price, even though no one opted into it.
You cannot turn off / opt out.
Except... if you open a support ticket.
Only then does Google enable the setting to turn it off!
I did:
More of a wishlist:
- Better media literacy.
- For Gen Z/alpha to be less afraid of talking to someone via phone/voice.
- shame-induced change in politics.
- higher taxes for corporations, the rich.
- hand-written letters between friends.
- more original films/IP
- longer attention spans
about a "society in which superstition defeats reason": "There are no checks and balances in a world where only charisma matters, no rule of law in a world where emotion defeats reason - only a void that anyone with a shocking and compelling story can fill." www.theatlantic.com/magazine/arc...
09.01.2025 09:00 β π 11 π 4 π¬ 1 π 0Feel like weβre living peak SF: seeing a tech worker bring their dog into work on a Nuna baby stroller.
07.01.2025 19:20 β π 1 π 0 π¬ 0 π 0
Survived three re:Invents so far. IYKYK.
06.12.2024 08:18 β π 2 π 0 π¬ 0 π 0
