𝚂𝚊𝚞𝚛𝚊𝚋𝚑 𝚁𝚊𝚒 ✦

Shipped a big update to 𝗥𝗲𝘀𝘂𝗺𝗲 𝗠𝗮𝘁𝗰𝗵𝗲𝗿. Paste any job description → get a tailored resume with matched keywords, cover letter, and email draft. Now you can stop rewriting your resume for every job application.

I see it in the top this month's too. 💖

One of my Midjourney's creation for v8.1 Made it to the top this week. Here's the prompt if you want to try it out: Illustration for a playing card. The card is the queen of fruits --profile yt39get --hd --v 8.1 --ar 2:3

@skeptrune You're absolutely right

imagine prompting codex from here

@thdxr wtf 😶

tired of this misinformation so we made a video on the truth behind the anthropic vs opencode drama

𝗙𝗢𝗖𝗨𝗦 How do you find the time to lock in and focus on the important parts of your work? Without distracting yourself from the calling of the AI tools? Likely, I've observed: • AI tools create a sense of feeling productive. • You try a lot of things, most of them in development, and ship nothing. • Day ends, repeat the same cycle tomorrow. (This is my case, because the amount of stuff I've generated is huge; if I'm able to ship 40% of that, it'll be huge. Share tips 🥺

@jpschroeder It's going to be fun, benching these models fr fr

We’re announcing: VibeBench, a new benchmark for what actually matters — how models feel when used on real work by experienced software engineers. But, we need your help. Here’s how it works: 1. An initial cohort of 1000 qualified software engineers (join: vibebench.standardagents.ai) 2. Groups of 250 evaluate new models for 2 days on real work. 3. Participants subjectively rank the model relative to other models they have experience with. 4. On day 4 a report is released with objective results derived from the subjective tests. How can you help: 1. We all need this benchmark to exist, but for it to become reality, we need an initial cohort of 1000 qualified software engineers. If that’s you, please join! vibebench.standardagents.ai 2. Repost this! We need to reach as many qualified engineers as we can find. 3. Share this initiative with everyone on your engineering teams. Together we can make this benchmark a reality for all of us.

💳If embedded finance is on your roadmap, this is 5 minutes well spent. We're building the State of Embedded Finance 2026 report — and we need your input. 🧵1/4

@krotenWanderung Vibe Evals should be a thing.

I have an impression that while search applications have really evolved, the evaluation of these applications lags (vibe evals are a thing?) But I truly hope I am biased. LLM-as-a-judge, for example...

@MicroLaunchHQ 3

How many products will you ship this year? No matter what.

Hi @sama , Can we have shared memory between ChatGPT & Codex, in between sessions? - Have a chat about a project - Codex can access that memory - We code and update the website, codex updates memory - And then in some other project, can we use the same learnings?

@skeptrune @benjaminshafii I have the data to back this claim 😭

@benjaminshafii "github stars don't pay the rent"

“a developer was discovered with almost 90% of his brain matter missing, while living an almost normal life after bragging about his repo having 20k GitHub stars.”

@orcdev Likely that I want this to work out. Because the limits on Claude Code is insane, hoping @OpenAI to do some magic with GPT-5.5

in 10 days we'll all ditch codex and claude code and live inside cursor + grok

@HowToAI_ "He argued that generative AI is fundamentally inefficient." What we need is more deterministic answers than non-deterministic, probable ones. Have more control over the AI model, maybe, LLMs are not the way forward.

Yann LeCun was right the entire time. And generative AI might be a dead end. For the last three years, the entire industry has been obsessed with building bigger LLMs. Trillions of parameters. Billions in compute. The theory was simple: if you make the model big enough, it will eventually understand how the world works. Yann LeCun said that was stupid. He argued that generative AI is fundamentally inefficient. When an AI predicts the next word, or generates the next pixel, it wastes massive amounts of compute on surface-level details. It memorizes patterns instead of learning the actual physics of reality. He proposed a different path: JEPA (Joint-Embedding Predictive Architecture). Instead of forcing the AI to paint the world pixel by pixel, JEPA forces it to predict abstract concepts. It predicts what happens next in a compressed "thought space." But for years, JEPA had a fatal flaw. It suffered from "representation collapse." Because the AI was allowed to simplify reality, it would cheat. It would simplify everything so much that a dog, a car, and a human all looked identical. It learned nothing. To fix it, engineers had to use insanely complex hacks, frozen encoders, and massive compute overheads. Until today. Researchers just dropped a paper called "LeWorldModel" (LeWM). They completely solved the collapse problem. They replaced the complex engineering hacks with a single, elegant mathematical regularizer. It forces the AI's internal "thoughts" into a perfect Gaussian distribution. The AI can no longer cheat. It is forced to understand the physical structure of reality to make its predictions. The results completely rewrite the economics of AI. LeWM didn't need a massive, centralized supercomputer. It has just 15 million parameters. It trains on a single, standard GPU in a few hours. Yet it plans 48x faster than massive foundation world models. It intrinsically understands physics. It instantly detects impossible events. We spent billions trying to force massive server farms to memorize the internet. Now, a tiny model running locally on a single graphics card is actually learning how the real world works.

The effects of sugar on your body and lifestyle are insane. I've been trying to quit sugar. Two days in. There are no crashes, likely used to feel sleepy and lethargic by afternoon, but not today. Having a high concentration and focus & no brain fog.

Soon, you will be able to vibe code accounting integrations... 🔹 Get all the invoices from Quickbooks, Netsuites, Xero, DualEntry, etc.. 🔹 Fetch all CRM data, find which are the active leads. 🔹 Fetch bank-feeds from Xero in real time. Do all this & much more, while vibing. ... soon. 😁

Try it: apideck.com/cli

Debugging a broken Quickbooks integration used to mean jumping between 4 tabs, Postman, and our dashboard. Now I just type apideck accounting get-invoices in my terminal. We shipped the @apideck CLI. It brings accounting, ATS, HRIS, and CRM data straight into your shell, and into Claude Code. One command. Real data. No context-switching.

@rxhit05 Somewhere between -- -idea in notes -landing page

What stage is your SaaS in right now? -idea in notes -landing page -MVP built -first users -making money where are you stuck?

@mattyp Prompt?

hues & views