aichannel

Back when Travis Kalanick was problematic for Uber, I made the conscious choice to try to only use Lyft. I did that same thing originally with OpenAI vs. Anthropic and used Claude for similar reasons but as it turns out Anthropic is mostly better anyway.

Writer. Building @pinatacloud. Working on simple static site hosting https://orbiter.host \ https://polluterofminds.com

ngl I think Anthropic is just as ideologically bad as OpenAI, and the only way to use their products is a probationary "this is too good to not use"

but the second they're outpaced, no reason to fund their market capture.

𝖒𝖆𝖓𝖉𝖆𝖙𝖚𝖒 𝖆𝖉 𝖆𝖘𝖙𝖗𝖆

imagine a public taxi/responder dispatcher that just works.

I agree. This is absolutely a public optics thing. Lyft, for example, used many of the same underhanded tactics as Uber but they didn’t get as much bad press

i dabble with ai using open data from crypto https://alexpaden.tech

agreed, though the meta is compensating with costlier methods.

my guess is that smol reasoning around simple GUI tools will yield the cost collapse that finally outpace proprietary anti-science model.

FOSS can vibecheck like none other.

it’s dei for ai models and sadly the propaganda mostly comes from elon being mad at sam

Ive never had consistently better answers from anthropic personally.

And grok isn’t open source so obviously that shit didn’t matter. I’m just glad I get to know how the models work generally, and I want even better ones.