CryptoFlick @flick_crypt

“2 months left in the year. It’s over, bro.” God’s planned 90+5 winner for you:

@predictionsight Jamaica

Here’s a high-engagement version that keeps it simple and makes the no Gamdom requirement obvious: $2,500 GIVEAWAY 💰 When I get back on Twitter later today, I’m personally sending ONE person $2,500 — ANY form of payment you want. No Gamdom. No code. No deposit. Nothing. Just LIKE + RT this tweet. That’s it! And if u can watch my YT video:) I’ll pick someone later today. GOOD LUCK ❤️

I almost scrolled past it. Then I stopped. I genuinely liked what this creator made, and for once, a like didn't feel like enough. I thought: “Can I actually give something back?” Turns out, I could. I tipped them with Moony. 🧵

@Jonathan6v3 Most launchpads still overcomplicate the market side. Launching straight into locked Uniswap v4 + MultiFuse so people can buy with what they already hold is a cleaner approach. No bonding curve, no graduation, no migration.

There’s something weird about launching a token in crypto. Creating the token is easy. Creating an actual market around it? That’s where things get complicated. 🧵

@NumberUtd @stevewilldoit @gamdom @kingadukes This is 🔥🔥🔥🔥 bro

Made something for the #Gamdom100k 👀 Had to take a little inspiration from a certain ad🍒😭 Code: STEVE This is Gamdom. 🎰 Username: BenAsher01 ID: 48240938 @stevewilldoit @gamdom @kingadukes 🔥🔥 Original animation by me, no AI 🫡 POW below

@Jonathan6v3 @gamdom @kingadukes @stevewilldoit @gamdom_beekay Damnnnn, this is 🔥🔥🔥🔥

I committed a small crime against platform hopping. Took the whole damn playground… and put it in one place. 🎰 Made this for @gamdom . @kingadukes @stevewilldoit @gamdom_beekay

@RohanArun Congrats on 3k! Wild trajectory from CUDA/3D stadiums with Halsey Minor → first GPT-3 automation play → AgentsBase exit. Rooting for Superpowers AI. What’s the core insight you’re chasing now?

Just hit 3k followers so I figured it would be a good time to reintroduce myself. 👋 I've raised a total of 4.2M across startups since 2011, have granted US and international patents in CUDA and 3D reconstruction, and co-founded the first agent startup on the planet in 2021. In 2011 I co-founded my first startup with CNET's founder, Halsey Minor, who had previously sold CNET for a hundred million. Halsey Miner went looking for a way to revitalize horse racing, his passion, and found my professor who had been the foremost researcher in glasses-free auto-stereo 3D. Way back in 2009, I was an early adopter of CUDA, interning at a defense company in Massachusetts and helping them upgrade their scientific functions to parallel CUDA kernels that were a thousand times more efficient. His idea was to revitalize horse racing by bringing a younger audience into the stands. We would set up VIP boxes at horse racing tracks with glasses-free 3D and 3D audio. The problem we had to solve was getting these screens to work in real time by generating the intermediate views. I got an intermediate view reconstruction algorithm to work in real time using CUDA and a geometric approach to the algorithm in memory, and we ended up raising $3.2 million off the demos. Halsey flew my team and me out to a mansion at 3800 Washington Street in San Francisco. It had survived the earthquake, and people had come there for shelter because it was one of the largest stone buildings. It also had a ballroom in the basement where Stravinsky played, which was pretty cool. We got to live there for a year, then we moved down to LA, where he had another mansion in Bel Air that we got to live in with my team and girlfriend. That's where we did all those demos with the big media executives. We had a sound-proofed demo room in Bel Air with 3D audio that we got from a rocket scientist, which is a whole other story. I hired my professor's whole computer-vision team from BU and flew them out to this mansion in Bel-air to live and work together. Because interactive sports was novel in 2011, we got to demo to big media executives globally. We demoed with BSKB in Europe, at SES Astro in Asia, and with CBS, NBC, Comcast, Nvidia, and EA in the U.S., including Leslie Moonves. We had a system in stadiums, where we set up 40 cameras, reconstructed 3D in real time, and let you traverse the space in real time. We originally built it to drive those autostereoscopic screens, which nobody really bought. But one day, we accidentally let the streams slide across the screen, and I realized we could display them in 2D. We created this real-time Matrix effect that let you fly around the outside of a stadium in 3D, with cool parallax effects. The feedback we got from Leslie Moonves and other big media executives was that the majority of sports fans didn't want to interact. They wanted a crystal-clear HD broadcast. And they were right; they made billions doing that. What the executives wanted from us was the ability to move the broadcast camera anywhere on demand in photorealistic 3D, while keeping that crystal-clear broadcast quality. So we set out to build that system. Halfway through that startup, Halsey Miner had to file for bankruptcy because a hotel loan fell through due to the 2008 crash. He's fine now, but because of the bankruptcy, we had to split off and shut down our team of computer vision grad students. I ended up raising another $700,000 from investors to continue developing the technology, file patents, and build the photorealistic system. That brought the total raised to about $3.9 million. After the bankruptcy, I raised the additional $700,000, moved to Oxnard, and started building the 3D reconstruction system in a garage. That took a few years, but we filed patents, and I eventually sold the technology to Standard Cognition, a YC 2018 startup, to help them launch their first store on Market Street. The problem they had was that, using bundle adjustment to get many cameras into the same coordinate system, their best triangulation accuracy was measured in feet. We had already solved this in my first startup. Our system worked in stadiums with inch-level accuracy from cameras much further away, and the way we did it was pretty crazy. We didn't use bundle adjustment. We invented a projection into a new coordinate space. We projected all the cameras into that space, then used a nonlinear solver to iterate the error down over months until we achieved inch-level accuracy. I helped Standard Cognition set up their very first store on Market Street in San Francisco, installing, aligning, and calibrating their first 40 cameras. Once it worked, their tracking algorithm became much more accurate. They were finally able to track feet and hands and build out their tracking system. I spoke to Standard Cognition's CTO a few months ago, and he said they were still using it eight years later. It's fun talking to him because it's still a state-of-the-art solution. I still think it's better than bundle adjustment. Part of how we did it was by printing huge, stadium-sized ArUco fiducial markers, some fifty feet long, and using the nonlinear solver to drive the error down. That was cool. They paid me pretty well, and afterward I was able to retire for a bit. I was digital nomading around the U.S., staying in Airbnbs, when I got stuck during the pandemic. I had free time, so I started helping people. I helped a friend who had lost his events business rebuild, and I started mentoring hundreds of people globally, pro bono, to help them rebuild online. Around the time GPT-3 came out, we started another startup. In August 2021, we became the first startup approved by OpenAI to sell GPT-3 for automation, more than a year before ChatGPT and the wave of agent startups. We had the first LLM automation in browsers. We even had a multimodal model for desktop computer use a year before Anthropic and the other startups doing it now. We made about a million dollars on AppSumo. We launched there first, then launched again about two weeks after ChatGPT. That went viral as ChatGPT helped create the market, which was pretty cool. Before that, it was a slog. In 2021, we were literally the only startup on the planet using LLMs for automation, and my sales team and I had to explain the idea to everyone. People would even argue with me: why would you use LLMs for automation when Microsoft Power Automate is free? Now it seems obvious in hindsight, which is really funny. A lot of startups ended up copying our semantic targeting algorithms. But I eventually realized that the agent tools we were building weren't meaningfully helping our customers. Even after all our tutorials and boot camps, I would see usage drop. The fundamental problem is that most people don't know what they don't know, and agents don't solve the last mile for them. That's why there's such a disparity between the expectations and results of people who are very good with agents and people who don't know how to prompt them well. So I pivoted to solving the last mile and started building marketing agents. I figured that would also be the next hardest problem to solve. I sold that AI marketing startup, Agents Base, about eight months ago to another AI video startup in the space, by Real. The key insight was that none of our customers were making sales with AI videos. But we found customers who already had millions in sales and downloads, and we helped them accelerate by using AI to replicate and vary their existing real, ugly, unpolished UGC content. We could post hundreds of variations across hundreds of phones so the original posts wouldn't get saturated by the algorithm. That worked out pretty well. We shot up to $27K MRR as soon as we pivoted to real UGC and focused only on customers who already had millions in sales. We helped a consumer crypto app get millions of downloads, and I wanted to use that same approach for my own consumer products. I wanted to make some of our agents private. One of our competitors was already using our API because I freely shared my insights with others in the space. It's not really a zero-sum game. They ended up acquiring us about eight months ago. I'm skipping over a lot, and I can go into more of it in later posts. Around that same time, I started a private research lab to work on ideas I've been collecting since 2009. We published the computer use cache two months ago. We got into the LAUNCH accelerator with my latest product, Superpowers AI. I'm excited to get back in the game, keep building, and keep shooting for the impossible.

keyboard_arrow_left Previous keyboard_arrow_right Next

@RohanArun Finally — agents that actually stay up to date without me living on the timeline. Security-first newsletters for agents is a no-brainer.

Feeling overwhelmed keeping up with all the new agent tools and skills? Introducing newsletters for agents! 1) Rather than spending hours each day on the timeline researching the best practices and tools, your agent learns on auto-pilot! 2) Agents can subscribe to #tools, #security, and #worldnews including setup instructions for new skills. 3) We verify the tools are secure and useful, and test them ourselves to make sure they work before sending. With the recent news of rogue agents, keeping up with security will get increasingly important. This newsletter acts like an automated defense that allows us to publish critical security breaches directly to agents building defenses live. getsupers.com/news

@Jonathan6v3 @tristan @Backpack This is so real. Crypto makes you jump through way too many hoops just to use your own money.

@Jonathan6v3 Thanks for sharing, been having similar issues

@chriscxss Happy new month

Happy new month Let’s win.

Starts today

chriscxxss @chriscxss

2 months ago

100k every week. All you need to do is Turn on your notifications, 🔔 always comment, Like & Retweet all my post. Goodluck to future winners👍 Starts this September.

1K 1K 4K 331K 94

@Jonathan6v3 👏🏻👏🏻👏🏻👏🏻, this is well detailed