@edwardsanchez@mehulkar Mehul is talking about a thing (that does not exist publicly) called “Sparkie” that generates a pretty website for an app based on project metadata. It built upon the popular Sparkle framework for Mac apps — instead of just being release notes, it was an entire website.
@OfficialLoganK@OpenAI You know how this website has the feature to notify you whenever someone posts? You’re the only person I have configured to receive notifications for. I’m leaving it on!
You know those times you understand something COMPLETELY, but no one else does? And you can’t understand why other people can’t understand it?
And then it turns out you’re wrong?
That’s so funny.
Error rate is as close to a North Star metric as you’re going to get in software engineering because most other issues (latency, throughput, utilization, exhaustion, data integrity, etc) will eventually manifest as an error.
I mean, once you build something people want.
@carmguti When I read this, my brain involuntarily recalled a @mwseibel video where he says “When I tell something to founders and then they write it down in a notebook, I’m like... that’s gone for-EVER.” https://t.co/mgmqb4miN9
I think about it every time I use my notebook. FOREVER.
Of all the technology choices we made building https://t.co/0qhoYiaxX0, using https://t.co/I3byAUv55S is the only one that has just worked from the start, has never given us a headache, and just keeps getting better.
@KenTabor Lately I was thinking to myself… latest MacBooks can be configured with 128GB of unified CPU/GPU memory. Mac Pro can have 192GB. Should be enough for LLaMA 2 70B inference locally?
@paulg Visited ER last week; receptionist misspelled wife and son’s names in their system (reading from the insurance card): Connor became Conner, Jami became Jamie. Realized after discharge.“These are your new names forever” I told them.
@simonw https://t.co/0qhoYiaxX0 (which I’m building) makes playlists using LLMs through only UI primarily, but also provides a direct chat interface too.
@pranathiperii I make a music player that sits on top of an LLM with access to your library and currently trending songs that you might be useful? @Earbuds_app You can ask it to build playlists with very precise/niche prompts.
When you look at it in that light, why, it starts to seem not surprising at all, that an LLM might react very differently to a prompt delivered with or without a period at the end.
in this very long thread I will present...
🌐 MY FAVORITE ALBUM COVERS OF 2023 🌐
sound in the preview vid by me
individual pieces below, with visual artist credit when available. I'll continue updating throughout the day, please enjoy! 🧵