one issue I find with letting agents use `deno eval` is that there are certain logging styles that prevents VSCode from capturing the terminal output and passing it to the agent? if I press Enter on the shell, it passes an empty result
kinda sucks that indo ecommerce has turned into "make a video unboxing the product with no cuts and editing, or no refunds"
someone needs to make another plex that doesnt suck
seems like a good article that actually explains what that chatgpt study was about [thebsdetector.substack.com/p/the-cognitiv…](
The Cognitive Debt of Digging Through Preprints
Your Brain on MIT Media Lab
 ) [The Cognitive Debt of Digging ...](
The Cognitive Debt of Digging Through Preprints
Your Brain on MIT Media Lab
 )
I like that almost every UI interaction in GitHub's pull request page is broken when a new comment arrives, and I'd then have to reload the page for it to work again
oh copilot premium request limits are in effect now huh
asking the agent to rewrite the interop tests into expect calls was a mistake https://pds.kelinci.net/xrpc/com.atproto.sync.getBlob?did=did:plc:ia76kvnndjutgedggx2ibrem&cid=bafkreigz2cwnrbwru2uzukj7nfe2bmyenh57pv7uaawoiaruaovgs3hn2u
ow managed to get copilot to send in a garbled PR comment
letting agents know they can do `deno eval` to test things out seems to help hrmmm
ww3 outfit ideas