Get live statistics and analysis of Dan Benyamin (Æ)'s profile on X / Twitter

🚨 First, the big news: A patent I authored has been granted, and is now the property of Aeon. The invention? A method to convert the web into a linear video experience, providing an endless stream of information at your fingertips, on any topic you can imagine. Now, let me share the backstory I've been thinking about AI video production for a long time. In fact, that patent was authored 15 years ago! Over the years, we have seen AI provide a number of useful, although fairly straightforward, value propositions. Companies like @OpusClip focus on making existing video production easier for professionals. Others, such as @LumenFive , make video generation more accessible for people who lack the skills or funds to hire a professional team. But at @get_aeon , we're tackling something different. We are tackling scale. While AI and software are great at lowering costs and barriers to entry, their true strength lies in the ability to do things 1,000 times over without getting tired. ✨ Millions of stories that have never been told ✨ Think about it: for every video produced, there are millions of pieces of content – written, photographed, or conceived – that never make it into video format. Imagine turning a hotel review webpage into a stunning travel video, a list of the best neighborhood burger joints into a mouth-watering visual experience, or an in-depth news story hidden behind a paywall into an informative trailer. Now, picture this happening for every hotel, every burger place, every news story – a million times over. That's our mission at Aeon. We've developed the most sophisticated pipeline to create videos from anything you can imagine at the scale and control modern media companies require. We're going from thought to visual expression instantly, at an unprecedented scale.

Hey @levelsio, I know you are heading to LA. I hope you have a well-versed guide out here. It's sort of the difference between these two:

It is absolutely a blood bath trying to run a business building foundational models. @googlecloud is selling Gemini pro for $0.002 per second of video. That's about $10 to analyze a full length movie and write a 150 page scene by scene synopsis. It's *great* to be the app business though!

Cc @luusssso

~~E-Commerce test with #nanobanana vs my custom solution~~ I am on a quest to nail product placement with AI! After many sleepless nights, I am proud to announce Aeon Detail. Aeon Detail is a image-image tool for the sole purpose of maintaining the fidelity of product images when placed inside of other environments: - Up to 100 Megapixels - Full original product fidelity is maintained - Occlusions, relighting, and more I compare it to chatgpt, nano banana, flux kontext, qwen edit, and runway. Ideally suited for packaged goods, but soon hopefully anything.

🧠 Self-learning image models? Continuing my fight against AI randomness, can we have black box models learn? Borrowing concepts from GEPA and other automated prompt optimization, but this time in VLMs. Example of using AI to apply a sample wood pattern to a given image. Its challenging because both the perceived color and the wood grain need to be correct. Traditional image analysis (FFT energy analysis for grain) and color theory (Lab color space) allow for a scoring function, and an evolutionary algorithm (the E in GEPA) walks through the space and steers towards the best (final) result. This is the level we bring to @get_aeon to make AI actually work ;)

🧠 We may have cracked the hallucination code... This product is real, but nothing else is. You now can have any model wear or manipulate your product in full detail, with any length video, in any language, for social media or your product page. Pretty much every AI model messes up product details. Brands change, logos change, text is unreadable, and more. This is especially true for video. @get_aeon is at the forefront of maintaining product detail in video.

Hey @replicate I've had a number of inferences stuck in the booting stage all morning. Please help! 😬 Cc @lucataco93 @bfirsh

🚨 First, the big news: A patent I authored has been granted, and is now the property of Aeon. The invention? A method to convert the web into a linear video experience, providing an endless stream of information at your fingertips, on any topic you can imagine. Now, let me share the backstory I've been thinking about AI video production for a long time. In fact, that patent was authored 15 years ago! Over the years, we have seen AI provide a number of useful, although fairly straightforward, value propositions. Companies like @OpusClip focus on making existing video production easier for professionals. Others, such as @LumenFive , make video generation more accessible for people who lack the skills or funds to hire a professional team. But at @get_aeon , we're tackling something different. We are tackling scale. While AI and software are great at lowering costs and barriers to entry, their true strength lies in the ability to do things 1,000 times over without getting tired. ✨ Millions of stories that have never been told ✨ Think about it: for every video produced, there are millions of pieces of content – written, photographed, or conceived – that never make it into video format. Imagine turning a hotel review webpage into a stunning travel video, a list of the best neighborhood burger joints into a mouth-watering visual experience, or an in-depth news story hidden behind a paywall into an informative trailer. Now, picture this happening for every hotel, every burger place, every news story – a million times over. That's our mission at Aeon. We've developed the most sophisticated pipeline to create videos from anything you can imagine at the scale and control modern media companies require. We're going from thought to visual expression instantly, at an unprecedented scale.

~~E-Commerce test with #nanobanana vs my custom solution~~ I am on a quest to nail product placement with AI! After many sleepless nights, I am proud to announce Aeon Detail. Aeon Detail is a image-image tool for the sole purpose of maintaining the fidelity of product images when placed inside of other environments: - Up to 100 Megapixels - Full original product fidelity is maintained - Occlusions, relighting, and more I compare it to chatgpt, nano banana, flux kontext, qwen edit, and runway. Ideally suited for packaged goods, but soon hopefully anything.

It is absolutely a blood bath trying to run a business building foundational models. @googlecloud is selling Gemini pro for $0.002 per second of video. That's about $10 to analyze a full length movie and write a 150 page scene by scene synopsis. It's *great* to be the app business though!

✨ HIDDEN MENU IN AEON Well, maybe not so hidden ha. We are trying out a bunch of new settings to help make your videos come to life, starting with a parallax effect. What it does: • Instantly animates your static images • Creates depth and movement with a simple checkbox • Utilizes advanced vision models to determine optimal animation Just check a box in the feature menu, and then every image will be automatically analyzed and animated. It uses vision models to figure out exactly how the depth map should be animated, you don't need to do anything. Check out how to do it below: 👇 Do you have lots of images you wish you could make a compelling video from? Send me a DM! cc @get_aeon

Just had a chat with a 'human' on @QuickBooks and got this. Short and street style🤦‍♂️

What racing a triathlon teaches you about #b2bSaaS Time for a 🧵