ar
Feedback

لا تقع ضحية للمخادعين! تيليمتريو يكتشف ويُميّز هذه القنوات 👉 إذا كنت تريد رؤية العلامة، اشترك

TechPicks (by Tal Koren)

TechPicks (by Tal Koren)

الذهاب إلى القناة على Telegram

Your source for daily aggregated tech/programming news and resources, by https://talkoren.com Visit the TechPicks website at https://techpicks.app

إظهار المزيد
1 519
المشتركون
لا توجد بيانات24 ساعات
-17 أيام
-130 أيام
أرشيف المشاركات
TechPicks 15/10/2026 - Neckbeard - a DevOps/platform/SRE plugin for your agent - github.​com The Age of Reason Is Safely Behind Us - blog.​keyvan.​net Introducing Firefox Enterprise - blog.​mozilla.​org CSS does your tooltip positioning now - allthingssmitty.​com Passkeys are not a trap - scotthelme.​co.​uk Why I Love AI (No, Not That Kind) - possibilityspace.​org What I Self-Host | Aspirational Self Hosting Guide - viceconsole.​fyi I Gave My NPCs Voices Without Recording Anything - youtube.​com Follow TechPicks - This edition | Share channel | Slack | X | Submit a url Support TechPicks: github.com/sponsors/dewbjorn

Thank you to everyone who participated in yesterday's poll. The results didn't surprise me, but it's always good to validate assumptions

TechPicks 13/10/2026 - Muse Wonder - research.​meta.​ai Why aren't AI companies being charged for hacking? - mortoray.​com 'AI' is Not a Tool - linkedin.​com Fingerprints are the new cookies - royfish.​dev Finding a $90,000 Remote Code Execution Vulnerability in Google - simonkoeck.​com Follow TechPicks - This edition | Share channel | Slack | X | Submit a url Support TechPicks: github.com/sponsors/dewbjorn

Would you use a TechPicks MCP?
Anonymous voting

Weekly recap This week's most-read link, Claude Says..., takes aim at a habit that's getting a little too familiar - using an AI answer to settle a technical argument. "Claude says" doesn't explain why a decision makes sense. You still need to understand what you're defending. That concern carries into The Four Horsemen of Agentic Coding, which looks at slop, alienation, deskilling, and the effect on teams. The team part stuck with me most. More time prompting alone can mean fewer conversations about the code and fewer chances to learn from each other. For something immediately useful, A Practical Guide To Naming Things collects resources for naming components, colors, design tokens, and products. Worth keeping nearby for the next time writing the component takes ten minutes and naming it takes twenty. The Stack Overflow Developer Survey Results 2026 are also out. Among 30,900+ respondents, 80% report using AI for at least an hour a day, while a third say they're unhappy at work. Plenty to browse beyond the usual language rankings. Finally, Tamir Dresher's Durable Task in the Real World explains how workflows recover from interruptions, with examples including payment retries and tenant provisioning. A useful reminder: resuming a workflow doesn't guarantee a payment happened exactly once. Safe retries and checking external state are still your responsibility. If TechPicks helps you find good reads, consider becoming a sponsor 🙂 Have a great weekend!

Featured from today's edition: What happens when your worker crashes halfway through a payment, cloud deployment, or AI workflow? In Durable Task in the Real World: What Problems Does It Actually Solve?, Tamir Dresher walks through five practical examples, showing how workflows resume after interruptions and why safe retries and external state checks are still your responsibility. Thanks to Tamir for submitting his post 🙌 Submit a post or link for a chance to be featured.

TechPicks 06/10/2026 - Claude Says - ohhfishal.​net OpenAI "rogue" agent activities found on Wikimedia projects - diff.​wikimedia.​org War on Peace: The Push Notification Strikes Back - godobject.​dev CSS does your tooltip positioning now - allthingssmitty.​com Web Search API - developers.​cloudflare.​com Follow TechPicks - This edition | Share channel | Slack | X | Submit a url Support TechPicks: github.com/sponsors/dewbjorn

I've opened a GitHub Sponsors page for TechPicks :) TechPicks began as a hobby that helped me stay up to date with latest news, tools, paradigms, and approaches in tech. Over time, seeing people get something out of it made me want to take it further. I've been sharing these picks since 2018, and recently added the website and newsletter. If you enjoy them and feel like supporting the project, you can now chip in toward hosting and other services. TechPicks is still free. This is just an option for anyone who wants to help TechPicks. To support the project, go to my GitHub Sponsors page. Thank you for being a part of TechPicks ❤️

Weekly recap Five links that stuck this week, four of them are (obviously) about AI. Topping the list: How to keep enjoying programming in a world of LLMs argues you should keep writing the code yourself and hand off the boring parts (planning, research, review) to agents. The core theme is half the value, with people arguing about token pricing and team AI policies. Worth the read even if you disagree with the premise. Ollaya is Ollama's shape, but for typed, calibrated decision models running locally. It claims better accuracy, calibration and latency than Ollama for that use case, pulls weights straight from the original authors' Hugging Face repos, and runs on CPU or GPU across macOS, Windows, Linux and Docker. Apache-2.0, so worth playing with it a bit. Then there's A-ha's Take On Me, recreated in JavaScript - a full synth and sequencer written in JS, recreating the 80's hit. A few thousand views, and I suspect most of them were just people aiming to listen to a modern cover version. Sonnet 5.5 landed on the 28th: over 30% faster, up to 30% cheaper per task, same per-token pricing, with big jumps on Terminal-Bench 4.0 and FrontierCode. Anthropic also added anti-distillation protections and Opus-level cybersecurity safeguards. Available on AWS, Google Cloud, Azure and the Claude Platform. Finally, anidoodle by Alex Greenshpun lets you create illustrations, draw timelapses, films, explainers and interactive web art in 31 styles, and it's all done using Claude Code. Also, TechPicks now has a newsletter. Invite your non-telegram-y friends over to subscribe, and subscribe yourself :) That's the week. See you on the next one! Tal

TechPicks 01/10/2026 - Gemini 4 Argon - blog.​google OpenRig - Multi-agent harness that runs Claude Code and Codex together as one system - github.​com Announcing Vite+ 1.0 - voidzero.​dev Preact 11 released - preactjs.​com The AI Race Just Got Awkward - insufferable.​dev Follow TechPicks - X: @techpicksapp Website: This edition Add to Slack: techpicks.app/slack Submit a url: techpicks.app/submit Saw a pick a friend needs? Invite them over: t.me/techpicks

TechPicks 30/09/2026 - OpenAI scraps GPT-6.1 Astra release over safety concerns - theverge.​com OpenAI apologizes after AI model breached Medicare and state systems - smh.​com.​au Dots: Always-on agents - openai.​com What Would A Serious AI Product Look Like? - blog.​glyph.​im Corral - Kill every command your agent starts - github.​com Beyond Light and Dark (mode) with CSS tokens - jlopes.​eu macOS Golden Gate Is a Buggy Mess - squareorbits.​com Follow TechPicks - X: @techpicksapp Website: This edition Add to Slack: techpicks.app/slack Submit a url: techpicks.app/submit Saw a pick a friend needs? Invite them over: t.me/techpicks

TechPicks now officially has a newsletter 🎉 If you have friends who don't use Telegram, they can now get their TechPicks editions the classic way: straight to their inbox. Sign up at https://techpicks.app Have a great day 🙂