Maksim Terekhin 🇺🇦

@0xAX Hope you get a great opportunity soon 🙏 Market is flipped, it seems, especially for remote positions. Good luck 👍

I’m considering my next career step and would be happy to hear about interesting remote backend/infra roles. My background is mostly in Elixir/Erlang/OTP, distributed backend systems, Kubernetes, observability, and telecom/AAA platforms. I’m especially interested in technically strong teams working on infrastructure, observability, databases, Kubernetes, or distributed systems. I’m based in GMT+5 and open to working with teams across compatible time zones. If your team is hiring for something similar, feel free to DM me or point me to the right person.

@GergelyOrosz Noticing colleagues leaving AI generated replies in Code Review comments. Similar feelings 🙅‍♂️

A person I know (and who is? was? a good professional al) left an AI-generated comment under my LinkedIn post. Full-on AI slop. I asked: why do this? He replied. It’s because of “engagement.” People are burning their profession al reputation, paying for AI tools, for nothing

@arpit_bhayani Re: cross-timezone. From my experience, most of the time it is a planning problem and organizational (ownership) problem. If it is a feature plan accordingly, allow for delay. Critical fixes should be paged/alerted as such thus bringing clarity.

If someone is waiting on you for a code review, that has to be your P0 task. Look, waiting on a code review is one of the most frustrating things because the person is literally blocked on you. It gets even worse when there is a time zone difference to deal with. I get it. You do not have malicious intent and are genuinely busy with something important. But still, I would say it is a prioritization problem. Most people treat code review as something they get to when they have a quiet moment, and that quiet moment rarely comes. Coming from my personal pain point, I would say treat code review as a high-priority task, not a background one. If your teammate has raised a PR and you are the reviewer, that is important work in progress sitting idle. Every hour it waits is an hour of 'blocked momentum' (yeah, fancy term). Also, it is okay to preempt your non-urgent work for it. This matters even more across time zones because a delayed review does not cost an hour; it costs an entire working day for the other person (ohhh, this used to be such a pain). So the next time you see a PR sitting there, wrap up the review, because it is not "someone else's work"; it is yours. Hope this helps.

@cnakazawa And then folks use skill to write a skill which is absolutely useless

I really don't get the hype about skills. They are just docs. Just write docs and ship them inside your packages.

@kellabyte I am following the HoloStore story all the way and I must say you are doing an incredible job! Thanks for your effort. It gives an energy to do research and learn from what you are doing! I thought of building something like ClickHouse but eliminating the need for HouseKeeper

I'm so sorry about all the HoloStore tweets lately. I've been struggling w/ burnouts & all the things I wanna fix, try, build have lived rent free in my head. I just don't have enough hrs in the day. I'm excited for the first time in years cuz I've become less of a bottleneck

@kellabyte Wonder when we are going to see HoloStore at jepsen.io ☺️

I've started measuring HoloStore against etcd the consensus datastore that powers Kubernetes I always make sure coordinated omission is correctly accounted for I also benchmark the client. Clients are part of ALL distributed system I also mark HoloStore shard split events

I don't need more software. Would appreciate the software I already have to work better, though.

Unpopular opinion: "AI makes everyone a developer" is true the same way "cameras makes everyone a photographer"

@warpdotdev @thorstenball @AmpCode Amp is great! But it is too expensive alone for personal use. Using it with combination of ClaudeCode

Why use @ampcode? It's kinda the Porsche of CLI coding agents: – Polished TUI w/ fullscreen rendering – Smart routing across frontier models – Prebuilt sub-agents: Librarian (research), Oracle (deep reasoning) – Pay-as-you-go pricing, no token caps What do you like about Amp?

the @multigres operator is now open source progress so far: ◆ Direct pod management ◆ Zero-downtime rolling upgrades ◆ pgBackRest PITR backups ◆ Observability with OTel tracing link to repo in thread

@PThorpe92 Interesting, have quite quite the opposite experience. Migrated successfully quite a few projects from legacy system to new

One of the most attractive use case for clankers is massive refactoring or stuff like extracting functionality from one larger feature-set to another branch.. Yet they are all horrible at this, which is sad cuz it's exactly the kind of programming I actually don't mind offloading

You’re all wrong. The earth isn’t flat or round. It’s fucked

@qatarairways is violation any possible due care requirements. Pushing our flight back 5 times in a row, not providing any reasonable rerouting options, offering refund only for which there is no options to fly. Unbelievable, and this is 5 star airline company? More like a fraud

@vineeth161 @qatarairways Yes, my is cancelled at 04:15 am. What time is your flight?

@materAwe @qatarairways Hi , is your qatar airways app showing that this flight on march 16 is already cancelled .. mine shows ON TIME

Fascinated by how AI coding agents unlock education and make many things approachable. Learned about the @PeerDBInc recently, was curious about deploying it to my k8s environment. Couldn't find k8s operator, built my own with @AmpCode github.com/Neurostep/peer…

@qatarairways @qatarairways My flight from Colombo is scheduled at 03:24 an March 4. I can’t manage via the app. The phone number is Sri Lanka forwards to email. I sent email yesterday and there is no reply. Which phone number I can use to reschedule my flight?

@qatarairways is dropping the ball hard to help people navigate the disaster. Our flight is scheduled on March 4 early in the morning. They asked not to contact before 48hrs. Contacting now – nothing works. Can’t change booking at all. This is disastrous.

🚨BREAKING: Microsoft Research + Salesforce just dropped a paper that should scare every AI builder. They tested 15 top LLMs GPT-4.1, Gemini 2.5 Pro, Claude 3.7 Sonnet, o3, DeepSeek R1, Llama 4 across 200,000+ simulated conversations. Single-turn prompt: 90% performance. Multi-turn conversation: 65% performance. Same model. Same task. Just... talking normally. The culprit isn't intelligence. Aptitude only dropped 15%. Unreliability EXPLODED by 112%. → LLMs answer before you finish explaining (wrong assumptions get baked in permanently) → They fall in love with their first wrong answer and build on it → They forget the middle of your conversation entirely → Longer responses introduce more assumptions = more errors Even reasoning models failed. o3 and DeepSeek R1 performed just as badly. Extra thinking tokens did nothing. Setting temperature to 0? Still broken. The fix right now: give your AI everything upfront in one message instead of back-and-forth. Every benchmark you've seen was tested on single-turn prompts in perfect lab conditions. Real conversations break every model on the market and nobody's talking about it.