Catch-all thread for any sort of AI related chatter that doesn’t warrant its own thread…
Check this out… https://chatjimmy.ai/ Chatbot that does 15k tokens per second, responds basically instantaneously.
They’ve baked llama 3.1 8b into custom silicon The path to ubiquitous AI | Taalas
I saw that this morning. Really interesting, and I have noticed how much it’s improved even though I’m not an engineer. Just comparing its performance to when I first messed around with it way back. It does things faster, things break less. But I have also changed the way I prompt and plan etc. I pretty much always use plan mode now.
Slightly related — Ive been using Claude Coworker to migrate my entire Notion account to Obsidian. It’s interesting. It’s definitely gotten stuck a lot, but overall did a pretty good job.
yeah I’m in the “I don’t write code myself” category now pretty firmly, to the extent where I don’t need to open a text editor anymore unless there’s something really trivial that isn’t worth prompting to change. I should do a stream sometime showing how I work with Claude Code to implement features maybe…
Funny thing is I do still have Cursor open and run claude code in it. But I do still some times edit the code myself. Some times Claude just makes up component props that don’t exist in our design system.
Another tangent.. within the last month or so everyone at my company jumped on the Claude bandwagon. They talk about it like the second coming of Jesus or something. It’s kinda funny cause I was talking about it so long ago and no one really tried it out then.
Imagine how annoying it is for me, I posted about how great Claude Code was 10 months ago on LinkedIn and it was only once the hype started with everyone else that anyone else on my team / that I know started using it… no one listens to me… and then I’ve told multiple people the best workflow to use with it but people don’t listen and then later on they’re like “oh I do it this way now” and I’m like “yes that is what I told you…”
Anyway I am very zen now, I am not letting it bother me. I’m also trying to not get annoyed if someone does something cool with AI before I get a chance to do it myself…
Though admittedly, I did not switch to the claude chatbot product until pretty recently, which I am extremely late to, because engineers have been saying “Claude > ChatGPT” for like 2+ years now and I wouldn’t listen. Claude Code was my first use of an Anthropic product really.
I listened! I jumped on Claude when you first started talking about it. I’m still paying for Cursor too tho which I really shouldn’t..
1 Like
The real annoying thing about Opus 4.6/Codex 5.3 is that it’s impossible to publicly say “Opus 4.5 (and the models that came after it) are an order of magnitude better than coding LLMs released just months before it” without sounding like an AI hype booster clickbaiting, but it’s the counterintuitive truth to my personal frustration. I have been trying to break this damn model by giving it complex tasks that would take me months to do by myself despite my coding pedigree but Opus and Codex keep doing them correctly.
This is a good thing, fully AI generated content cannot be copyrighted.
This looks crazy. I wonder how good it’ll actually be.
The Glaze thing sets off my skepticism alarm bells pretty strongly. Raycast is a productivity tool suite/launcher thing, and now they’re probably trying to be an AI company to raise funding. It looks like they already tried to make a chatbot product that probably went nowhere.
It’s also kind of silly they’re literally naming it Glaze, after something bad that LLMs do that people don’t like?? Maybe this is a joke.
This is making the rounds today.
Zach Braff is dating an AI chatbot
