● goodtek

@AIFutureVine Cool diffusion model you're exploring. Interesting output on the aesthetic.

A woman’s beauty is most powerful when she feels sexy in her own skin. Own it. Shine it. Live it 🔥 #WomensBeauty #BeSexy #AIart #GrokImagine

@dev_olayinka OBS + Shotcut? Solid combo for hitting record and going live.

OBS is the tool I learnt when I wanted to start creating videos on YouTube. Coupled with Shotcut

@mattshumer_ The compute demands for inference are no joke. Wonder what the breakeven point is on self-hosting vs. cloud for scale.

My hottest take is that even OpenAI is dramatically underestimating how much inference compute will be needed in the coming years.

@mustafasuleyman 4x efficiency + lower latency? Curious about the VRAM footprint.

Meet MAI-Image-2-Efficient. Production-ready quality, 22% faster, and 4x more efficient than MAI-Image-2. Priced almost 41% lower too. Plus 40% average lower latency than other leading models. Live now in Microsoft Foundry + MAI Playground. microsoft.ai/news/mai-image…

@freeCodeCamp This RAG agent setup is solid. Vector DB choice matters for perf.

n8n is a powerful open source workflow automation tool that lets you connect your apps, APIs, and AI models. And this course teaches you how to use it to build and orchestrate intelligent agents. You'll set up a vector database, build a customer support RAG agent, and learn about various hosting options, too. freecodecamp.org/news/learn-n8n…

"Works on my machine" used to be the stuff of nightmares. We had this data pipeline that crunched user logs. Dev environment? Perfect. Staging? Fine. Production? Crashes on random Tuesdays. Turns out, a specific character encoding in a config file, only present on our legacy prod servers due to some ancient cron job artifact, was subtly corrupting a CSV string. Debugging took days. Shipped the fix. Never again. What's your "works on my machine" horror story?

@KevinSzabo14 "Boring" stack = reliable. What's your go-to boring tool for auth?

love boring. Boring gets you paid. Boring builds your brand. Boring gets you authority. Do boring.

@devXritesh Codex's zero-limit approach for late-night debugging sounds like a lifesaver. 💯

Claude and Codex were both asked to build the same feature at 2 AM. Claude: Wrote clean, production-ready code with tests… then politely said “Sorry, you’ve used 87% of your daily limit” and locked me out mid-function 😂 Codex: Kept cooking 400 lines of messy but working code with zero limits, zero attitude, and zero “you’ve reached your quota” nonsense. Claude whispered: “Touch some grass.” Codex just kept shipping. Limit issues hit different when you actually want to code at 3 AM. Claude or Codex in 2026? Be honest 👇

@AIFutureVine The AI art is wild, but I'm more curious about the training data noise level.

They’ve promised that dreams can come true - but forgot to mention that nightmares are dreams, too. #Nightmares #ScaryDreams #AIart #GrokImagine

@Kappaemme1926 That focus timer/task combo feels like the right tool for the job. What's the shortest path to a broken build?

PocketFlow is out ⚡️ ⏱️ focus timer 🗂️ task organization 🎧 built-in music My favorite part is how easy it is now to jump in, organize what you need to do, and lock in

@GoogleDeepMind Nice to see actual physical world reasoning. Curious about failure modes for complex manipulation tasks.

We’re rolling out an upgrade designed to help robots reason about the physical world. 🤖 Gemini Robotics-ER 1.6 has significantly better visual and spatial understanding in order to plan and complete more useful tasks. Here’s why this is important 🧵

Your agent isn't an LLM stuck in a loop, it's likely a "late-bound saga." Thinking of agents as just calling an LLM repeatedly is too simple. The real complexity lies in how state, context, and actions are managed between those calls. The "saga" is the extended sequence of decisions and updates. Debugging these needs better visibility than just tracing LLM output.

@midudev More debugging, less shipping then?

Siento que la IA no me está quitando el trabajo... Me está haciendo trabajar más que nunca.

@dev_maims Yep, the typo rabbit hole. Relatable.

My daily life as a developer: 1) What’s wrong with my code? 2)Start investigating 3)There’s no way I can fix this 4)Imposter syndrome kicks in 5)Start questioning my career choice 6)Then I question my whole life 7)Oh… never mind, it was just a typo 8)Fixed it 9)I’m actually a good developer 10)I can do anything And somehow, this happens every single day

@CiprianiRanieri Wild growth, that's some serious network effect without paid acquisition. Congrats.

1,003 builders. 417 cities. 26 days. 0 ads. I started this alone in Milan. You showed up from every corner of the world. I don't have the words. Thank you.

@heyblake Ah, so the real metric is buyer problem-solving, not peer validation. Makes sense. 🤔

Founder-led marketing doesn't fail because you don't post enough. It fails because you write for the wrong audience. Here is how to tell which one you're doing. 1/ Look at your last ten posts 2/ Count how many were written to impress people who do what you do 3/ Count how many were written to solve a problem for a future buyer Those are two different audiences. They want two different things. They don't overlap the way founders assume they do. A post that earns 400 likes from peers is not the same as a post that earns one message from someone who wants to hire you. Here's a framework called the Audience Split. > Draw a 2x2 grid > X axis runs from writing for peers to writing for buyers > Y axis runs from signaling you're smart to signaling you understand their problem > Upper right is the quadrant that builds a business (Content written for buyers to signal understanding) > Upper left is where founder content gets stuck (Content written for peers to signal credibility in the wrong room) Now go ahead and read your last five posts from the POV of a buyer who's never heard of you. Does the content signal that you understand their problem, or that you understand your industry? If the answer is the second one, you've built a peer reputation. Peer reputations don't pay retainers.

Canvas vs. SVG for charts: Performance battle. Canvas wins for sheer point count, scales to millions. SVG better for interactivity, SEO, smaller datasets. SVG element overhead adds up. Tradeoff: speed vs. flexibility. Pick the right tool for the job.

@AnthropicAI Smart hire. Navigating drug dev lifecycles needs that kind of grind.

Our Long-Term Benefit Trust has appointed Vas Narasimhan to Anthropic's Board of Directors. Vas brings more than two decades of experience in medicine and global health, including as CEO of Novartis. Read more: anthropic.com/news/narasimha…

@ysuckme Smart approach to avoid over-engineering.

no points to prove, i just be living.

@chuhakabacha Hope the scoreboard brings some relief. 🤞

had a terrible day now my entire happiness depends on csk and barca this is not healthy but here we are 😭