N

@xanf_ua Думаю, и сам наблюдаешь, что происходит не уравнивание, а наоборот, еще большее разделение.

Неделю назад было 7 лет, как я работаю в GitLab, так что была пятиминутка рефлексии Помню на собеседовании в GitLab мне задали вопрос - вот у нас есть люди, которые коммитят в webpack (это было 7 лет назад), есть крутые эксперты по a11y и т.д., а какую ценность вы как инженер принесете в нашу команду? Тогда я ответил что я, со своим большим бекграундом в DevOps привнесу во frontend-часть практическое понимание что за процессы и какие задачи стоят за всеми покрашенными кнопочками, что редкость среди адептов HTML и CSS В 2026 этот вопрос “чем вы выделяетесь?” становится актуальным как никогда. Пару столетий назад на Диком Западе великим уравнителем стал Кольт, а у нас великим уравнителем стали LLM. Теперь знание и даже практический опыт React, Next.js или еще чего-то не выделяет вас среди множества кандидатов. Почему к примеру я должен выбрать именно вас?

The relevance is highly affected by install vs uninstall rates, same as your organic ranking in the app store. So basically if you dropped it at some point, you'll need to pay a ton to try to recover. Upside is that if you'll get it higher, your organic ranking will be better too. Though this system is even more funnier considering that for popular apps that already rank good the ads would be much cheaper than for the others.

I will never understand how the Shopify app's relevance score works. I am sure there is a bug, but I can't prove it 😅 The app is basically a "combined listings" app. It is specified in name, in description, basically everywhere, but Shopify ads algorithm says it is not a relevant keyword 😅 This makes it inefficient to run ads in this case.

@xanf_ua @x1unix Если сравнивать с отдельным полным тайпчеком? Интересно, как тогда решаешь проблему перфоманса

@x1unix Есть интеграции, хотя мой опыт и замеры в опенкоде показали что интеграция с LSP даёт противоречивые результаты как минимум в typescript

Последнее время окончательно мигрировал с Claude Code на pi.dev везде где можно (к сожалению некоторые заказчики хотят только антропик по легальным причинам) Конечно же если вы хотите использовать Антропиковскою подписку - это возможно нелегально :) Основная причина - с ним суперкруто, благодаря RPC/JSON режимам строить воркфлоу разработки адаптированные под проект. Я прекрасно понимаю коммерческую выгоду условного claude-code/codex - возводя ИИ-агента "в центр" процесса разработки - мы провоцируем тратить больше токенов делая операции которые легко могут быть детерменированными через агента + МСР (или агента + CLI - в данном случае неважно). Pi же помогает мне генерировать сфокусированные решения которые ПРОЩЕ дебажить и которые фактически работают согласно закону Конвея отражая оргструктуру проекта

@therealchreke so, economically speaking, the migration is as efficient as firing the guy who suggests it?

@minofopts @ThePrimeagen @tsoding @ChShersh i feel like sqrt(-1) having repos in both rust and go

@antirez @KrishanghArjun never heard of it in mine, although a literal translation would make sense within a suitable context, yeah. in mine it’s more common to hear “it could be worse”. in this case, i guess, cultural differences do matter 😅

@KrishanghArjun @n7olkachev What about this Sicilian way of saying: "It can't be darker than midnight", to say that if it is a very bad moment it will likely improve. Does it exist in the same formulation in your mother language?

Today I discovered that in China there is this concept of "losing face" (mianzi / 面子), which maps 1:1 with Sicilian "perdiri la facci" both as exact wording and conceptually.

not really sure what you imply? if you mean that lots of different cultures just saw the same thing and came up with the same non-obvious idiom meaning the same thing, it would be even more surprising, at least to me. shared root makes more sense, although to me it still feels fascinating, some of them existed long before our languages themself and still exist, even with the cultures being so different nowadays.

@n7olkachev @antirez Crocodiles actually do the tears while eating thing so its unsurprising that's used all over

@trq212 @nateparrott will the LSP tools be back in the next versions?

the VSCode extension in particular is goated, a lot of @nateparrott magic in this one: x.com/nateparrott/st…

We also shipped a ton of new Claude Code features today, including: - checkpointing - a new VSCode extension - tab to think - a new mascot, we call him Clawd 🦀

@claudeai do not recommend updating: the new version seems to not have the LSP tools available, making it pretty much unusable.

We're also releasing upgrades for Claude Code. The terminal interface has a fresh new look, and the new VS Code extension brings Claude to your IDE. The new checkpoints feature lets you confidently run large tasks and roll back instantly to a previous state, if needed.

Introducing Claude Sonnet 4.5—the best coding model in the world. It's the strongest model for building complex agents. It's the best model at using computers. And it shows substantial gains on tests of reasoning and math.

@theo Writing scripts for a hugely unprofitable streaming service is not the flex you think it is.

the topic is extremely subjective and personal, but i don't think it's a good idea to convince yourself into "this is enough", it's hard to be confident you won't regret about that later on. i mean, if that was actually enough, would you even have such thoughts in the first place? wasted potential and talent could hit really hard. there's nothing bad with trying new things and going further either + usually it's pretty easy to return to where you were. imho it's more risky not to take action and then regret it.

From time to time we get criticized for making "yet another to-do list" product. Or a chat product. Or a messaging product. Or something we've kinda sorta already made before, just in a different form, combination, or approach. "How about something else? How about something bigger? How about something completely different?" @dhh even reflected on it personally in his recent @lexfridman interview: QUOTING--- We think of Elon as finding great talent, and I’m sure he is also good at that, but I also think that this beacon of the mission. We’re going to fucking Mars, we’re going to transform transportation into using electricity, we’re going to cover the earth in internet is so grand that there are days where I wake up and go like, “What the fuck am I doing with these to-do lists?” Like, “Jesus, should I go sign up for something like that?” --- So it's not just other people levying the charge — it's us too! I've given it similar time and similar thought. But then I step outside tech for a moment, ground myself in other professions, and realize how odd it would be to ask the same kind of question to a knife maker. "Hey master bladesmith, you’ve made a dozen knives this year. Same as last year. Same as the last 20. If you love metal so much, why not help build a battleship? Think of all that metal!" Or a cabinet maker. "Hey, ok I get it — you take raw wood and you turn it into beautiful furniture. Again and again and again. Don't you think a wood addict like you should consider getting into forestry? Imagine!" That would be a ridiculous line of questioning. So why is it persuasive in tech? I think part of it is that we're talking about software. An endlessly malleable medium that can truly be anything. With no conceptual limits, it’s easy to think you should keep expanding. But why? There's nothing at all wrong with honing in, developing your craft, making variations of things you're good at, and getting better each time. Nothing small about it. Nothing unfulfilling about it. So instead of looking sideways at what others are building — or upward toward the mythical “next level” — we focus ahead on what we're good at. We like to make useful, straightforward things we need. Specific tools and familiar ingredients combined in different ratios, different molds, for different purposes. Like a baker working from the same tight set of pantry ingredients to make a hundred distinct recipes. You wouldn't turn to them and say "enough with the butter, flour, sugar, baking powder, and eggs already!" Getting the same few things right in different ways is a career's worth of work.

complexity first, simplicity second people say “keep it simple,” but most approach it backwards. they start from simple, then add on complexity without seeing the whole. that’s how you end up with frankenstein products: clean-looking components awkwardly stitched together, held in place by duct tape and wishful thinking. true simplicity emerges only after you’ve grasped the full complexity first. you can’t abstract away what you don’t fully comprehend. once you deeply understand the entire system — the edge cases, feedback loops, emergent behaviors — then the elegant patterns start to surface, creating solutions that genuinely click. people often misunderstand complexity as the enemy of simplicity. but complexity isn’t the enemy, it’s reality. your goal isn’t to ignore complexity, but to master it. when you think holistically, you create systems whose parts reinforce each other rather than clash. the UI naturally mirrors the underlying data model. the API aligns seamlessly with how users think. the entire product feels inevitable. real builders dive into the messy reality and embrace it. they map out the bizarre edge cases, user mental models, technical constraints, and business pressures. they sit patiently with complexity until the right patterns emerge. only then do they craft the simple, intuitive interface that makes all that complexity invisible. it’s like a swan, serene on the surface but paddling like hell beneath. this is why Notion succeeds where most productivity apps fail. we didn’t start by saying, “let’s build a simple notes app.” we asked, “how would people organize and share information, with the fewest primitives” then we built abstractions that aligned with those conceptual models. systems thinking is essential because it’s the only path to building products that scale — not just technically, but cognitively. users shouldn’t need to grasp your internal complexities to extract value. that’s the paradox: the more deeply you embrace complexity in your thinking, the simpler the experience becomes.

so disappointed to see these recent bugs with BFS status (we also got an incorrect “security incident” email not so long ago). with the 30% rev cut, weren’t we promised with better quality?

anyone sells digital subscription products in EU? there seems to be a bug with taxes calculation, where the total price isn't differ between countries (although the displayed price on PDP is different and seems to be correct). the problem seems to be specific to subscription products.

Toki's first Customer Hub implementation is live! Really enjoyed working with Solly Baby's team on this!

ah that's the reason of the sudden traffic pike i got 😅 funny to see automated accounts getting much more traction than the original posters, but i guess it is what it is

lol yeah a great one indeed

@sashhedgesky cursor + sonnet 4 does insanely good job for me rn. i'm not sure whether it was cursor v1 release or sonnet 4, but i'm really happy with this setup.

@n7olkachev last time i had claude write me integration tests it deleted my production database 🙃

lol claude got so excited after 10 minutes of fixing tests