Gleb Reys

great news

You know what? After trying to make @openclaw working on older Intel Macs and later remote Linux server… I think people do have a point about dedicated Mac mini! Looking forward to massive productivity gains soon :)

@_architected This is the way

ставишь omarchy на dell xps - отваливается сеть после ребута - чинишь сеть - замечаешь что не работает звук - чинишь звук - ебешься с конфигами для внешнего монитора и внятного скейлинга и переключения между окнами хайперленда - заебало ставишь ubuntu - инсталлер сломан на моменте выбора тайм зоны ставишь cachy os - вроде бы все ок, но он почему то подтянул новую версию KDE в которой не работает половина плагинов вот и день пролетел. ultimate linux experience.

@Palak3312 It's much harder to fix hardware or recover after mistakes

Hardware engineering or software engineering — which is harder? 👇

@Saint_n0mad That’s a mighty microscope! Great setup - I have temporary setup but no storage drawers so for now it’s tricky to organise the space

Invest in home lab, you won’t regret.

Ok last few upgrades complete broke the @openclaw for me. Think we need the emergency mode in openclaw doctor where it’s reviewing things critically with the recent upgrade in mind. Also, probably need something like “upgrade notes - check/fix these first” suggested by openclaw update

awesome toolkit of CLI capabilities ready to integrate with OpenClaw

@Merocle Is it €25K in one stack, or more? Wonder how much this is in Anthropic Opus tokens. Really looking forward to using one of these soon!

"Free tokens"

@_architected @sudakov_maksims а почему удобнее заметки на нём, чем на iPad? стилус какой-то особый или просто экран приятный?

@sudakov_maksims Мне нравится делать заметки на нем, куда удобнее чем на айпаде. А так помимо чтения как бы функциональности не сильно много – книги читать удобно, мангу и ПДФ тоже.

Приучил себя вести рабочие заметки на киндле и получаю от этого какой то непередаваемый детский восторг. Из минусов, постоянно хочется отвлечься на книгу.

NotebookLM and recent integration/sync into Gemini make this even stronger (actually, NotebookLM is where/why I got upgraded to Google AI Pro)

@kverlin @_architected а я в последнее время полностью ушёл в заметки Raycast. оказалось, что главная польза от Bear была таки в удобстве и скорости создания заметок, а не собственно в чистой-красивой презентации.

@_architected У меня оба стоят и я чаще пишу в Bear.

кастомизированный Obsidian может выглядеть не хуже чем Bear, мне нравится.

No way! Was talking to Gemini on my way to work earlier today, looking for exactly this functionality. Thanks

Didn’t realise there wasn’t a native app. Should try soon

@michalmalewicz Great idea! Looks awesome as well

This. I am building this. Custom, no GUI typewriter with e-ink, raspberry PI and a 2015 apple keyboard. Wrapped in black birch plywood, with brass plate at the bottom. Will post updates as I go. Yes, the golden sheet under the components is the brass already! :-)

@Saboo_Shubham_ It's almost like each release now needs "do these steps if you're upgrading" - because lots of new things are showing improvement only if you activate them

OpenClaw with GPT-5.4 feels noticeably better with the new update. GPT-5.4 (thinking=high + fastmode=true) seems like the sweet spot. Better judgment, less back-and-forth, still fast enough for daily use.

@shuravko Сразу видно - серьезный тип :) Я тоже теперь такую хочу

Подумал с утра, почему у меня нет кепки как у Томаса Шелби из Peaky Blinders? И срочно исправил. Благо в Лондоне можно легко купить именно эту модель этого производителя, за что и люблю сей город.

OpenClaw v2026.4.11 just dropped 🦞 This one is all about memory. You can now import your ChatGPT history and turn it into real, structured knowledge. Plus upgrades across chat, video, and integrations. Here’s what actually matters ↓ 1. ChatGPT import → Memory Wiki This is huge. Import your past ChatGPT conversations and convert them into structured memory. • Organized insights • Source tracking • Browsable history New “Imported Insights” and “Memory Palace” views let you explore everything inside the UI. Your agent now builds on your past. 2. Webchat feels way better Chat is now cleaner and more structured. • Media shows properly • Voice replies look clean • Embed support for external content Everything feels more polished. 3. Video generation upgrade Way more flexible now. • URL-based assets • Reference audio support • Better aspect ratios • Higher input limits Agents create better videos without the old friction. 4. Feishu and Teams upgrades Feishu: • More context • Reactions • Typing indicators Teams: • Full reaction support • Better pagination • Easier OAuth setup 5. Plugin and Ollama improvements • Plugins can define their own setup flows • Faster Ollama model discovery Much easier to extend and scale. 6. Fixes that matter • OAuth issues fixed • Audio transcription working again • macOS mic flow fixed • Better WhatsApp and Telegram handling • Improved timeouts and failover • API edge cases cleaned up No major breaking changes. Mostly smoother and more reliable everywhere. How to update openclaw update openclaw –version npm install -g openclaw@latest Docker: pull latest image and restart This one is a big step. Agents are starting to remember your entire history.

@iamlukethedev Great stuff! And just the right time because I’ve been considering doing the opposite - looking for ways to convert OpenClaw memories into something I can share or use with Claude Desktop or ChatGPT

At the end of the day, @AnthropicAI have a great product with the biggest complaint for it being (un)affordability for edge cases like OpenClaw… This is a classic product placement challenge and nothing a company this big can’t solve in the near future. Bright future but hope they change the open source sentiment for the better, soon.

@steipete @openclaw Keep working sounds like an implied instruction when you delegate something to agent

Two experiments in the next @openclaw to address some "GPT is lazy" issues: 1) Strict mode: agents.defaults.embeddedPi.executionContract = "strict-agentic" This tells GPT-5.x to keep working: read more code, call tools, make changes, or return a real blocker instead of stopping at “here’s the plan.” docs.openclaw.ai/providers/open…