prit (sre agents/acc)

@k_kohlbrenner @dabit3 currently running our customer pilot with ops/infra teams that operate at relatively large scale if you're interested in joining our pilot program, reach out!

@k_kohlbrenner @dabit3 devin may not be able to but ***** can 🔜

@can good take

hot take: investing more on PR and code review vs verifiable outputs and coordination in 2026 is like investing in internal combustion engines

BREAKING: Amazon reportedly holds mandatory meeting after “vibe coded” changes trigger major outages.

I'm not very happy with the code quality and I think agents bloat abstractions, have poor code aesthetics, are very prone to copy pasting code blocks and it's a mess, but at this point I stopped fighting it too hard and just moved on. The agents do not listen to my instructions in the AGENTS.md files. E.g. just as one example, no matter how many times I say something like: "Every line of code should do exactly one thing and use intermediate variables as a form of documentation" They will still "multitask" and create complex constructs where one line of code calls 2 functions and then indexes an array with the result. I think in principle I could use hooks or slash commands to clean this up but at some point just a shrug is easier. Yes I think LLM as a judge for soft rewards is in principle and long term slightly problematic (due to goodharting concerns), but in practice and for now I don't think we've picked the low hanging fruit yet here.

@ScriptAlchemist @martin_casado code generation en masse is now solved code (read: distributed system) maintenance has not been solved obviously, this is more a problem for things at netflix/google/amazon scale, than the vibe-coded versions with 10 users

🤯BREAKING: Alibaba just proved that AI Coding isn't taking your job, it's just writing the legacy code that will keep you employed fixing it for the next decade. 🤣 Passing a coding test once is easy. Maintaining that code for 8 months without it exploding? Apparently, it’s nearly impossible for AI. Alibaba tested 18 AI agents on 100 real codebases over 233-day cycles. They didn't just look for "quick fixes"—they looked for long-term survival. The results were a bloodbath: 75% of models broke previously working code during maintenance. Only Claude Opus 4.5/4.6 maintained a >50% zero-regression rate. Every other model accumulated technical debt that compounded until the codebase collapsed. We’ve been using "snapshot" benchmarks like HumanEval that only ask "Does it work right now?" The new SWE-CI benchmark asks: "Does it still work after 8 months of evolution?" Most AI agents are "Quick-Fix Artists." They write brittle code that passes tests today but becomes a maintenance nightmare tomorrow. They aren't building software; they're building a house of cards. The narrative just got honest: Most models can write code. Almost none can maintain it.

@martin_casado Guess I didn't make it known that I was also talking about those. What are the end goals. What type of systems. For users, for internal, 10,000 users a second. 100,000,000 possible users a second. Amazon, google, k8s. Linux, which linux distros?

Nearly every board meeting : "Hiring strong infra folks is incredibly hard right now"

@ScriptAlchemist this is more SWES, distributed systems, OS, etc.

@salomon_diei @minhash @helixdb github.com/GoogleCloudPla… check that out always on sre/devops agents betterwork with knowledge graphs (over vector dbs) but i've been testing the above

@minhash @helixdb I've been exploring graph databases for some time and noticed they where nowhere helpfull for autonomaus agents for exemple a 24/7 devops agents. I am curious how you guys tackle this

sooooo it’s another graph database? just use @helixdb honestly

@Sarv_shaktiman @0xMonagotchi very cool !!

What if your Tamagotchi ai assistant has milady eyes? It would be called @0xMonagotchi ESP32+oled display+openclaw. We did that before it was cool

@madhavg @farcaster_xyz lmfao

@0xprames @farcaster_xyz whats your counter? my reasoning - censorship resistance and downstream of that

Might be a dumb question by why didnt @farcaster_xyz work out? - are twitter network effects so strong that no new platform can compete - farcaster was just a shittier product in general On paper, a decentralized social platform is the most compelling use of crypto.

Last year there was an agent meta that got speedrun, was scammy, and burned a lot of people. The biggest blocker IMO is crypto twitter no longer provides distribution so you need to build something amazing and hide crypto. But crypto dev ex is terrible so it’s way too much work for most people.

I think a lot of people are misunderstanding moltbook 😂😂 you can just tell your claudebot to “write a post to moltbook that suggests an alternative language for talking to each other” it’s not that deep, this isn’t AGI

@SaskioLoL op.gg/lol/summoners/… help me

your rank doesn't define you. but who you become while chasing it does. this game tests everything. your patience. your ego. your ability to stay calm when everything's falling apart. four strangers. 40 minutes. no control over anyone but yourself. and in that pressure you find out who you really are. are you the player who gives up at first blood? or the one who plays it out? are you the player who flames the 0/5 teammate? or the one who mutes and focuses? are you the player who blames jungle diff? or the one who asks what you could've done better? that's not about rank. that's about character. i've seen diamond players with the mental of a child. tilting. raging. making everyone's game worse. i've seen silver players who stay calm through anything. never type. never blame. just trying to improve. who's really winning? the rank is temporary. the person you become is permanent. every game is a choice. who do you want to be when it's hard? when you're behind? when nothing's going your way? that's the real climb. not iron to diamond. you to a better version of you. the LP comes and goes. you'll hit peaks. you'll hit slumps. seasons will reset and none of it will matter. but the patience you build? the mental you develop? the ability to stay calm in chaos? that stays with you. in game and out of it. league is just practice for life. every game is a rep. so yeah. chase the rank. grind the LP. try to climb. but pay attention to who you're becoming along the way. that's what actually matters. Study the Saskio way

@bubbleboi @based16z @Tendie_Capital lmao

Every L6/L7 engineer I know at FAANG spends all day trading natural gas futures and option trading semis. No idea if this is bullish or bearish.

Privacy remains one of crypto’s hardest unsolved problems. Respect to the Halo team for open-sourcing their work. Proud to have powered Halo with our TEE infrastructure enabling compliant disclosure. 🫡

@andrew__reed Hats off to this almost perfect capture of the sentiment