swyx

swyx

@swyx.io

i want everyone to aspire to integrity, intensity, and intentionality. see also @latent.space

9.1K followers151 following276 posts

swyx reposted

If you're using SvelteKit remote functions, you might want to patch asap. @swyx.io did a neat TLDR on HN: news.ycombinator.com/item?id=4663...

news.ycombinator.comall DoS attacks and one XSS. this isnt as bad as the react server components CVE... | Hacker News
Svelte@svelte.dev

We've released fixes for 5 CVEs affecting the Svelte ecosystem. Please upgrade your apps! Read the post to learn if you're affected: svelte.dev/blog/cves-af...

swyx reposted

Following on with more info for @swyx.io @mackuba.eu maintains sdk.blue The ATProto Touchers Discord discord.atprotocol.dev is great for all the newest launches & troubleshooting. Soon @slices.network will be a couple of clicks to make a new backend & aggregator for ATProto data.

sdk.bluesdk.blueA curated list of libraries & SDKs for the Bluesky API and AT Protocol
Boris@bmann.ca

Hey @swyx.io! Yeah, it's just a pile of tech talks, the conference & some events, and it's me and a handful of other folks from the #ATCommunityFund, not a developer hub! There are no paid devrel or ecosystem roles by anyone yet! Lots more community infra needed.

swyx reposted

swyx reposted

Fascinating to hear evolution of behavioural engineering in the context of generative AI models. Always useful to hear what @swyx.io is up to and their views. The Agent Reasoning Interface: o1/o3, Claude 3, ChatGPT Canvas, Tasks, and Operator —with Karina Nguyen of #OpenAI www.latent.space/p/karina

latent.spaceThe Agent Reasoning Interface: o1/o3, Claude 3, ChatGPT Canvas, Tasks, and Operator — with Karina Nguyen of OpenAIKarina Nguyen from OpenAI (and previously Anthropic) discusses her work on Claude, ChatGPT Canvas and Tasks, and the new AI interaction paradigms for human-computer collaboration.

swyx reposted

hi friends, our handle is now @latent.space! so nice that @substack supports bluesky! we'll be looking to publish more here going forward. automation tools and ideas for content u like to see welcome

if you only learn 1 thing from philosophy 101 i hope it is deontology vs consequentialism

*chinese models cannot say what happened on june 4 1989* “reee they are so censored dont trust them” *openai cannot say david mayer* “haha so cute silly baby anyway thanks for working on agi for the good of humanity”

hot take: this is not any better than getServerSideProps direct db access in server components is great dx for hacky starters and then you’ll want to undo it more and more as your app grows and needs o11y/security/auth/caching etc theres one killer app for RSCs - reducing js bundles. stick to that

virgin ai researchoor: > argmax benchmarks for bold rows > suffix CoT model with “o1” to game lemmings > y axis starts at 70 > begins prompt paper citing vaswani et al > github only has readme.md chad ben affleck: > llms are library of vectors of meaning and transformers that interpret in context

> low key reminds us he made Good Will Hunting with a buck fifty in late fees age 24 > describes Succession fanfic where Kendall and Stewie elope > AI will replace DVDs > refuses to elaborate > is best batman