HackerLangs
TopNewTrendsCommentsPastAskShowJobs

ntcho

10 karmajoined 5 anni fa

Submissions

OpenCode Data: Real-world AI model usage, cache ratios, and costs

opencode.ai
2 points·by ntcho·3 giorni fa·1 comments

comments

ntcho
·3 giorni fa·discuss
Came across this dashboard tracking live LLM usage on OpenCode users. Interesting production metrics, including token volume, session costs, and cache hit ratios. DeepSeek v4 currently dominates their market share.
ntcho
·11 giorni fa·discuss
Gosh, the last sentence is the most terrifying thing I’ve read today.
ntcho
·mese scorso·discuss
Average Firefox experience right here
ntcho
·2 mesi fa·discuss
absolutely love the domain here. great taste
ntcho
·2 mesi fa·discuss
I landed on a similar workflow. I found letting agents navigate / self-update docs on a repo full of markdown files quite successful
ntcho
·2 mesi fa·discuss
username checks out
ntcho
·2 mesi fa·discuss
Have you tried MacsFanControl? MBP fans rarely turn on by themselves, but controlling yourself makes throttling not a problem
ntcho
·3 mesi fa·discuss
I’m also having a blast with generating userscripts/css to make web SaaS work better for my workflows. Went from waiting for months for a new feature to writing QoL improvesments myself
ntcho
·4 mesi fa·discuss
I also noticed Gemini web UI failing to send messages on Firefox. Anyone else?
ntcho
·6 mesi fa·discuss
I think this should be on HN hall of fame: https://hn.algolia.com/?q=sabotage+manual
ntcho
·8 mesi fa·discuss
Decades of progress made in hardwares and networks… just to be negated with 100kb+ webpages
ntcho
·9 mesi fa·discuss
This is hilarious, thanks for sharing. Kinda crazy how well it works and already better than some ads
ntcho
·9 mesi fa·discuss
Has some quirks, but Beeper is exactly what you are describing: https://www.beeper.com/
ntcho
·9 mesi fa·discuss
This reminds me of cleaning a toaster with a dishwasher: https://news.ycombinator.com/item?id=41235662