Digital nomad at SEA š Building the web by day, experimenting with sound by night. āØš¹ šš¤šššš
Phuket via Tampa
Joined February 2022
- Tweets3.8K
- Following392
- Followers260
- Likes11K
future job title: data producer at web
As Anthropic prepares to IPO for over $1 trillion, a reminder of Dario Amodeiās words shortly after founding the company:
āDistributed human labor is being used to train AI models by large, centralized actors, who concentrate the resulting profits while in the long run making the human labor obsolete.ā
In other words, this IPO will be built off the backs of the worldās creatives - who will get nothing out of it except obsolescence.
It is a totally unjust transfer of wealth and opportunity from artists, writers, musicians & other creatives to a tiny tech elite.
I hope it is remembered as such.
(this quote is from a document he wrote that was unsealed in Bartz v. Anthropic)
google ai answers is just terrible. it oftentimes gives answers based on wiki and fandom, such as this one where they donāt even give visible credit to fandom for the wrong answer.
saw you were primarily building from mobile now. liberating, isnāt it? wonder how much longer people will call it coding.
what are the odds?
anyone seen chatgpt give a chat a name with chinese characters from english-only source? just gave it an @ABC7 article and asked for more info. weird...
šÆ like an extra set of eyes. and doing taxes with them is great when the goal is reducing investment cost basses.
and when claude detects shitās going down it warns users about mental health to help avoid illness due to stress. crazy. wouldnātāve believed it two years ago.
crazy how claude can take a doc and know itās a scam right away. last year i fed a lease into claude after signing and it told me to hire a lawyer pronto, then laid out the whole scam.
donāt trust. verify.
one stopping point was claude asking if it could add a lib called cf-wasm/og and a rust tool called satori so it could make open graph social sharing cards for the site it built.
the library had less than 300 stars on github but claude surfaced it anyway. github.com/fineshopdesign/cfā¦
tested opus 5.5 on long horizon tasks by feeding it a 20 page claude design doc, a reference implementation, and a prompt. I told it to work as autonomously as possible and it worked the better part of 4 hours alone and only stopped to ask me a few questions.
tied chatwise.app from egoist last year but never found myself building a habit around it. do you tend to use many model providers?