spruce_tips

5 ms·

1.
by spruce_tips 8d ago
what counts as discounted rate plans? if i pay for a year in advance (and get the yearly discount) and have train on my data set to off.. are you saying that is still being trained on?
2.
by spruce_tips 7mo ago
Yea I knew embeddings I just didn’t quite understand it in OPs context. Makes sense, thanks
3.
by spruce_tips 7mo ago
sorry - what do you mean by embeddings in your last sentence?
4.
by spruce_tips 7mo ago
https://github.com/JoeBerg8/tollbooth this was part of a little saas tool i was building (since retired it) so spent some time today having an LLM help me pull it into a headless service. far from perfect but sharing a
5.
by spruce_tips 7mo ago
https://github.com/JoeBerg8/tollbooth this was part of a little saas tool i was building (since retired it) so spent some time today having an LLM help me pull it into a headless service. far from perfect but sharing a
6.
by spruce_tips 7mo ago
https://github.com/JoeBerg8/tollbooth this was part of a little saas tool i was building (since retired it) so spent some time today having an LLM help me pull it into a headless service. far from perfect but sharing a
7.
by spruce_tips 7mo ago
no one paid me but didnt really have this running for very long on my inbox. was really just a poc. and you're right - the human side is weird. surprisingly hard to solve the "real human, not spam, that's also an email addres
8.
by spruce_tips 7mo ago
I built a side project to solve this for myself that’s basically an inbox toll system. It funnels emails from unknown senders into a hidden mailbox and auto replies to the sender with a payment link. After the sender pays, the email gets r
9.
by spruce_tips 7mo ago
Opus is a superior brand line to Sonnet because historically it’s been a more powerful model. I think the thinking behind a rebrand is that people wouldn’t have as willingly switched their usage over from opus 4.5 since that model has been
10.
by spruce_tips 7mo ago
Ah yes, explains why it takes 3 seconds for a new chat to load after I click new chat in the macOS app.
11.
by spruce_tips 8mo ago
Tried it, not impressed. Terrible UX and generally just confusing. Didn't really intuitively know where to go and why. The thing that made me most mad was the clunkiness around pulling specific files into the chat input as context. Lik
12.
by spruce_tips 8mo ago
https://read-faster.com/read/SgIcbUqJ
13.
Read at 600+ wpm using rapid serial visual representation (RSVP)(read-faster.com)
2 points by spruce_tips 8mo ago|0 comments
14.
by spruce_tips 8mo ago
wow, it's almost like this makes sense
15.
by spruce_tips 9mo ago
composer 1 has been my most used model the past few months. but i only use it to execute plans that i write with the help of larger, more intelligent models like opus 4.5. composer 1 is great at following plan instructions so after some car
16.
by spruce_tips 9mo ago
it took me a lot of tinkering to get this feeling seamless in my own apps that use the api under the hood. i ended up buffering every token into a redis stream (with a final db save at the end of streaming) and building a mechanism to let c
17.
by spruce_tips 10mo ago
try java 25, and update your priors :)
18.
by spruce_tips 10mo ago
it's in cursor only
19.
by spruce_tips 10mo ago
no, i dont think it's a heavier lift to test different model families. my point was that swapping models, whether that's to different model families or to new versions in the same model family, isn't straightforward. i'm
20.
by spruce_tips 10mo ago
I agree there is no moat to the mechanics of switching models i.e. what openrouter does. But it's not as straightforward as everyone says to switch out the model powering a workflow that's been tuned around said model, whether tha
21.
by spruce_tips 10mo ago
100% agree. composer-1 really has been the sweet spot for me of capability, reliability, and speed. i dont ask it to do too much at once, and this approach + its speed, materially speeds my work up. i generally find i get the most out of mo
22.
Ask HN: Is anyone using Devin/Cognition?
1 points by spruce_tips 11mo ago|0 comments
23.
by spruce_tips 1y ago
yes, in a software engineering context, always use ask mode instead of agent mode unless youre truly doing dumb, tedious work that youve done many times before
24.
by spruce_tips 1y ago
lol
25.
by spruce_tips 1y ago
Wonder if deprecating direct access means the gpt5 can still route to those behind the scenes?
26.
by spruce_tips 1y ago
These presenters all give off such a “sterile” vibe
27.
by spruce_tips 1y ago
Building low cost tools to give email owner's their attention back. 1. Inbox Toll - solves the "real human, still spam" problem by making the sender pay a fee before the email actually lands in your inbox. Set your toll as lo
28.
by spruce_tips 1y ago
checkout java 23. it has come a long way
29.
Ask HN: Persisting LLM token streams through a page refresh?
1 points by spruce_tips 1y ago|0 comments
30.
Ask HN: Persisting LLM token streams through a page refresh?
1 points by spruce_tips 1y ago|0 comments

More ›