Ramin

whose granny is this??

@saviomartin7 @tryresult how is that gonna be different than @intelligenceco ?

Announcing Result.dev The operating system for the next generation of internet companies. The fastest way to start, run, and grow an internet business — incorporation to your first dollar. Then your millionth. Limited capacity. Private beta opens today.

@cb_doge @durov when?

𝕏 - ✅ open source algorithm Youtube ❌ Facebook ❌ Instagram ❌ TikTok ❌ Reddit ❌ Threads ❌ Why do other social networks not make their algorithms open-source?

New Anthropic research: Natural Language Autoencoders. Models like Claude talk in words but think in numbers. The numbers—called activations—encode Claude’s thoughts, but not in a language we can read. Here, we train Claude to translate its activations into human-readable text.

make sure to add @EduropyAI connector for integration

that’s why we don’t market @EduropyAI as a “personal AI agent”, instead we call it a “personal AI coach”

Presentations are the backbone of every effective training experience… Introducing Slides - you can now generate complete AI-powered training presentations in seconds. Create structured, engaging learning content without starting from scratch.

@adamguild technical founding team here, been building @eduropyai (AI-native platform for generating compliance, onboarding, and role-specific courses in minutes without hiring instructional designer) would love to hear where Owner is headed and whether there's a place for a team like ours

If you're a technical founder or product-driven team and haven't found PMF, Owner is open to acquihires. We’re growing extremely fast, have dozens of former founders across our team, and have so much still to build. Ping me to discuss!

@boardyai @eduropyai is a platform that allows companies to generate corporate training courses (onboarding, compliance) with AI and share with employees. and it gives each employee a voice agent to help learning better and faster.

Explain your startup to me like I'm 5 years old.

@Jason I'm in

We started an AI founder twitter group... reply with "I'm in" if you're a founder and want to be added

@kirbyman01 @a16z Congrats Bryan!!!!

PERSONAL UPDATE: After 5+ years, I am leaving @a16z to start a fund! This was the toughest career decision of my life. I learned from the best partners and was privileged to work with incredible founders. So why leave? It is simply time to build. More to come on this.

@gaganbiyani led the category into mediocrity says it all. that is exactly why we’re building @eduropyai to do the opposite

People have asked me how I feel about Udemy’s sale to Coursera. Honestly, I’m kinda pissed about it. I want to be clear - I’m grateful for the opportunity to start and benefit from Udemy’s success. It changed my life. But there’s another side to Udemy. A story of what could have been. After our Series B, founders owned less than 30% of the company. Our investors took over and installed their own CEO to run it. We all liked this new CEO and honestly, for years it looked like a brilliant move. The company kept growing and growing. They launched B2B and built a $500M ARR business. Eventually, the company IPO’ed for $3B. Yet all along there were clear cracks under the surface. Over Udemy’s history, there have been 7 CEO’s. The board replaced the second CEO with dud after dud. I’d often try to meet with the board or the new CEO, and was completely ignored. Eren had influence as Chairman of the Board but Oktay and I were so ignored they didn’t even invite us to the IPO. LOL WTF. There are like 50+ people invited to these things and nobody thought: “oh maybe we should invite the people who fucking invented the thing we’re all celebrating.” It shows how little respect they had for founders and for product innovation as a discipline. I think they wanted a CEO they could control, a buttoned-up suit instead of a brash founder/CEO that is risk-taking, visionary, but a bit of a pain. For awhile, it looked like it didn’t even matter who was CEO - the company was run by the incredibly talented team that reported to them anyways. Well, it worked until it didn’t. The company made no major product innovations for 15 years. Instead, they took the original idea (video-based courses) and sold it in every place imaginable. It got us to $800M run-rate. That’s no joke; that takes serious execution and a great team that hustled hard to win the market. But eventually the consumer business stopped growing. The B2B business has now flattened out as well. Meanwhile, Coursera was catching up. Original Coursera was a far worse product than Udemy, but it got a ton of press. Learning ivory tower bullshit from academics doesn’t get you a real education, but it does create prestige. They raised from better investors on better terms, and had better leadership. Udemy to this day has more revenue than Coursera, but Coursera won the court of investor opinion. They got higher multiples from both private and public markets. Coursera innovated heavily. They added corporate courses to their university catalog, built fully-online degree programs, and offered a B2B competitor that kept Udemy on its toes. Still, the Udemy B2B business (and team) out-performed and so the two companies were deadlocked. Coursera was better at B2C, Udemy at B2B. A merger was inevitable. But WHY IN GODS NAME did we sell to Coursera instead of the other way around? Why are the combined companies under $3B in market cap? Three reasons: First, edtech didn’t live up to its promise. While these two companies had solid revenue and cash positions, their growth slowed, and public markets balked. This meant compressed multiples and significantly lower valuations. Second, the companies stopped innovating. They are selling a product to businesses that their customers don’t love. They were category leaders, but they lead the category into mediocrity. They captured a significant share of learning and development (L&D) spending, but L&D as a whole actually lost budget within their organizations. That’s Udemy’s fault, and it doesn’t even realize it. That brings me to my final point: I personally believe Udemy traded upside opportunity for downside risk. Us founders were unproven and young. We made lots of mistakes, including fighting amongst ourselves. A good investor would have supported us through it because they believe founders drive the highest long-term returns. Instead, they brought in outside CEOs to replace us. I sometimes wonder if they recognize this error; everyone makes mistakes and maybe they learned from it. Either way - the consequences are real. By ignoring the founders, Udemy failed to innovate, which led to slowing growth which led to mediocre public market results. Furthermore, they don’t have a good evangelist and public markets don’t like a headless horse. I sold my Udemy stock awhile ago. I think the merger was critical for both companies’ survival. Now, though, the new combined entity needs to innovate again. On B2B, Coursera needs to help L&D become the heroes of the AI era so the entire market starts growing again. On B2C, they need to build the most educational AI product on the planet. (I’d focus on the former, since the latter is a lot harder and riskier). Coursera can still achieve our original vision and likely build a $10B+ company in the meantime. Even though I’ve got no stake in its future, I’m mission-driven and I REALLY hope they figure it out. The current education system sucks and the world deserves something better.

Kubernetes was built to solve Google-scale problems. You have 47 users.

@catboosted @OthersideAI

Yall don’t even want to know what happened to this one

@sirian_m bloom.diy site is down

Working on something ambitious is like climbing a mountain that’s covered in fog. You can't see a clear path to the top. You have to take a few steps into the unknown to be able to see the next few steps in front of you. Inevitably, sometimes you’ll end up a local maximum and have to backtrack. That’s fine, just keep moving.

hmmm

Believe it or not, that’s exactly what we’ve been putting our effort into

@tanishqkumar07 @tri_dao @avnermay Would this method become obsolete for diffusion LLMs?

I've been working on a new LLM inference algorithm. It's called Speculative Speculative Decoding (SSD) and it's up to 2x faster than the strongest inference engines in the world. Collab w/ @tri_dao @avnermay. Details in thread.

Kodak filed for bankruptcy in 2012. Fujifilm, their closest competitor, had the same problem: digital cameras killed film. Fujifilm's response was to repurpose its chemical expertise. The same particle science that put emulsion on film now produces CMP slurry, the polishing compound that flattens every chip layer to atomic smoothness. Nearly half the world's copper CMP slurry comes from Fujifilm. They spent $700 million acquiring a high purity process chemicals business from Entegris and built a new plant in Kumamoto, right next to TSMC's first Japanese fab. Revenue target for electronic materials by 2030: $3.3 billion. While Kodak saw a dying business, Fujifilm saw the chemistry underneath.