Профиль

I was skeptical about @sifa.id because I hate LinkedIn but seeing that it has lexicons for talks has sold me on it... I joined, registered everything and spent a bunch of time on a relay that allows me to use it to update the talks from @sifa.id while keeping ricciuti.me/speaking fully prerendered 😎

ricciuti.meTech blog and personal website of a mad scientistricciuti.me

Let's say I connect to jetstream hosted by bsky and I put myself in the wantedDids...if my account is on a different PDS will I get any event?

I changed my phone cover and I swear it is faster

⣿⣿⣿⣿⣿⣿⣿⠿⠛⠛⠛⠛⠿⣿⣿⣿ ⣿⣿⣿⡿⠟⠋⠀⢀⣠⣤⣤⣄⡀⠀⠻⣿ ⣿⡿⠋⠀⢀⣤⣶⣿⣿⡿⢿⣿⣷⡄⠀⢹ ⣿⠁⠀⣴⣿⣿⠿⠋⠁⣀⣀⠉⠻⠃⠀⢸ ⣿⡀⠀⢿⣿⣷⣤⣾⣿⣿⣿⣿⣦⠀⠀⢿ ⣿⠃⠀⠈⠙⠿⠿⠟⠋⢁⣼⣿⣿⠀⠀⢸ ⡇⠀⠀⣿⣶⣆⢀⣤⣾⣿⣿⠟⠋⠀⢠⣿ ⣿⡄⠀⠙⢿⣿⣿⣿⠟⠋⠀⢀⣠⣶⣿⣿ ⣿⣿⣦⣀⡀⠀⠀⠀⣀⣠⣶⣿⣿⣿⣿⣿ ⠛⠛⠛⠛⠛⠛⠚⠛⠛⠛⠛⠛⠛⠛⠛⠛

So Bsky doesn't seem to pick up my @standard.site theme...is there a tool to check which theme combinations are allowed?

Sol was tasked with a decently simple task (medium complexity at most), but I asked to call Fable on loop until Fable found no issues...it ran for six hours. Agent review is nice and an almost required step, but a feel like they are often overzealous and report something minimal as critical.

OMG that "Ready for Compaction" feel so much like someone accepting their own lobotomy with a heartfelt last sentence that I almost felt bad sending the compaction command the second later.

I have to say that all in all the new version of MCP looks good. I just wish they had a better back compatibility story.

Wtf Kimi K3 on a MacBook with 32gb of ram is possible? Painfully slow, with 4k context but wtf

Yeah AI might kill us all but in the meantime we get a lot more Rob Miles video so I'll take that

Join me for today's episode of This Week in Svelte! Live on youtube.com/live/1EIIRwU... in 15 minutes or so!

This Week in Svelte, Ep. 148 — ChangelogYouTube video by Svelte Societyyoutube.com

no react is faster than #noreact

Is there some benchmark that pinpoints the context length at which a model start to degrade in quality? I keep compacting at around 100/200K because I have the feeling that even models with way larger context window start suffering at the range, but maybe I'm just shooting myself in the foot.

Lol

Would you look at Sol proposing a final review from Fable before declaring everything done...such a senior mentality!

I threw Fable at the problem with multiple reviews from Sol...now I have to review 70 changed files and 1000s of loc, around 200$ worth of tokens. I started reviewing it and the code is so bad I'll probably just scrap everything and start again with more incremental steps. "lol" but also "fuck" 😵

On the 25th of August I'll be in Hamburg to talk to TechCamp Hamburg and we are organizing an impromptu svelte meetup with @hmans.dev if you are in town or near it join us for some svelte chatty evening!

@theosteiner.de is the GOAT and @danabra.mov doesn't need presenting...go watch it to learn about at proto!

Omg today is a new release of the MCP spec and I have absolutely no will to completely refactor tmcp to fix all the breaking changes while keeping back compatibility. 😱😭😱