Sabitlenmiş Tweet
I've verbally accepted an offer to join as staff eng @buildwithfern to sling some yaml (the currency of the future).
I'm excited to be a part of a small team scaling it up with @getpostman.
English
cbro
4.8K posts
cbro
@broady
yaml/md/blob engineering @buildwithfern, former bigtech worker. coding for fun.
Katılım Nisan 2007
483 Takip Edilen3.1K Takipçiler
@jhleath Nice! How are you doing multi-writer with this architecture?
English
excited to announce serverless, multi-writer SQLite for Archil users.
even though file systems are the primitive for AI agents, builders know that some data is better represented as database tables. good news is that those databases run on our file system natively.
today, we're making it simple to add infinite SQLite databases to your file system, send queries/mutations to it from multiple locations, and only pay for the storage that you're actively using.
excited to see what you build!
Archil@archildata
Agents need structured data too. Developers are using SQLite to store agent history, state, and memories, alongside Markdown. Today, we're announcing serverless SQLite on Archil. Create unlimited databases on your disk, and use our SDK against them -- from multiple clients.
English
I get rules are rules but these visitation rules in the hospital are extremely ass for people with big families, no lie. 3 people ain't enough.
English
Over the past long weekend I almost started something like this - github.com/spf13/go-skills
Thanks @spf13 !
English
Don't care + Didn't watch + I can run MoEs 8 times the size of my GPU
Theo - t3.gg@theo
I'm sorry, I need to crash out about local models for a bit.
English
this could have been an email
and it could have stayed in drafts
Debasish (দেবাশিস্) Ghosh 🇮🇳@debasishg
Andrew's thoughts on the Bun Rust rewrite .. andrewkelley.me/post/my-though…
English
@Aitch_El @MitchPls4Real It's disrespectful to the animals, I have no issues with eating animals I do however have a problem with them being slaughtered for nothing. >_>
English
I worked nights at Woolies during uni.
There were always stacks of perfectly good rotisserie chickens and fresh bakery items left at closing.
At first, night staff could take them home free or heavily discounted. Then head office banned it - no markdowns, no staff takeaways.
Every night I’d throw 20+ good chickens and piles of bread in the bin.
One coworker took a couple home for her kids. She got caught and was fired.
They wouldn’t even donate to charities or food banks.
talegio difficult #boycottMSO 🇸🇩🇮🇷@talegio212959
65c off and expires today. Who are they kidding
English
Yeah so the move to Slack is going well
dax@thdxr
someone at @SlackHQ please add a reply feature. i'm begging you just copy discord DO NOT GET CREATIVE please
English
@cunnyphile1488 @chrisczynski @kerckhove_ts had to explicitly disable single process because it leaked every new tab and window on my Arch system. I'm not blaming ghostty for this, but it wasn't a good experience
English
My buddy made a 1 terapixel rendering of the Buddhabrot fractal. He believes it to be the largest static render ever made.
2^20 pixels square.
English
this is actually a really big deal
especially for frameworks / ISR type workloads on top of workers
Brendan Irvine-Broque@irvinebroque
Workers Cache — Your Cloudflare Worker now has its own tiered cache in front of it: blog.cloudflare.com/workers-cache/ Workers Cache is a programmable tiered cache between Workers — even between code running in the same Worker we’ve seen pretty much every approach to caching over the years from our customers, and Workers Cache is the direct result of everything we’ve learned it takes all the workarounds and clever hacks we’ve seen to make sites and APIs fast, and simplifies them down to writing JavaScript and returning Cache-Control headers you can use for so much — for us this is like shipping 12 different things on the same day first up — you can use it to place your app near your database, cache its responses, and run code in front close to users:
English
No good engineer will tell you it’s useless to read the code btw
That’s the easiest signal
Gustavo Valverde@GustavoValverde
@zeeg Some say it's between reading and not reading code
English
Initial impressions using Fable 5 (High) so far
1. It's still consistent when it passes 200k context window.
2. It has a tendency to automatically commit on certain decisions, even though I told it a few times to give me its takes.
3. Building on the previous step, it's pretty good for long running sessions. It rarely needs human intervention.
4. I love how it explains things concisely.
5. It found so far a few outstanding edge cases that are very subtle in our operators. Those are well documented in our codebase, but It was fascinating to see when it pulled out those pieces for me.
I like it for designing and planning so far, and it's superb in that regards. I created several follow-up plans, each independent of each other, and if I want, I could just parellize them via sub agents (assuming they each work in their respective worktrees).
English
@ivanburazin @lox @renehernandezio If you think Kubernetes implies EBS then something has gone wrong in your evaluation process.
English
English
There is no category where Kubernetes wins for agentic workloads.
Performance: sandbox (90ms vs. 2s per instance)
Ergonomics: sandbox (2 commands vs. 12 steps)
Statefulness: sandbox (local NVMe vs. mounted EBS)
Spiky workloads: sandbox (2.7x more efficient)
Steady state: sandbox (2% more efficient)
So if you need a sandbox, use a goddamn sandbox.
English
@ivanburazin @lox @renehernandezio Is this really a comparison of a truck vs car?
Comparing k8s and a "sandbox" is more like a comparison between an interstate road system and a car.
English
@lox @renehernandezio Sure you can. But all that adds more complexity and cost - not less.
You’re trying to use a truck when you need a car.
English
in some personal news: after 10+ years here in the US, i’m now a permanent resident 🦅 🇺🇸
what an incredible wave of relief to have read “case approved”.
English
what is going on here?
GitHub@github
We heard you. And we agree. In light of recent developments in physical media, GitHub is proud to announce that you can now obtain your public repo on CD-ROM. Keep it. Lend it to friends. Pass it on to your children. Your code is physically yours, forever. Until you lose it, let's be real. Order yours today. gh.io/cd
English
Now that we have Fable back, let's see if we can tell
I gave all the Claude models (haiku 4.5, sonnet 5, opus 4.8, fable 5) a tweet-sized task
Can you spot the Fable?
Steve Krouse@stevekrouse
I used Fable nonstop while it was out and am now back to Opus, and I don't notice a difference If you told me that there was a bug with my Claude Code for those three days, and I was on Opus the whole time, I wouldn't be surprised I am skeptical about all the claims of how much better people find it. Not because I don't think it's better. I trust Anthropic evals. But because I think our guts are poorly calibrated to sense differences in intelligence at this level My guess is that it's a lot like blind taste testing of wine: it's orders of magnitude harder than you'd think it is. It's easy to fool yourself that you can tell the difference Which I guess we can turn into a challenge: when Fable comes back online, I can make a "blind taste test" app to give people a chance to see if they can tell which is which. I'd be very impressed with those that can! I'd love to learn your ways!
English