This week in AI

Welcome to Synthetic, the newsletter on all things AI. If you like what you read, please consider sharing this with others who might find it useful.

Open AI in on a Hollywood charm offensive before launching its video generator

It’s widely reported that this week Open AI will be doing the rounds in Hollywood, promoting its new Sora video platform. It’s already in the hands of some A-list directors and actors, and when Tyler Perry saw it demonstrated he immediately cancelled his plans to build an $800 million production studio. Change is a-coming, and there will be a lot of crying before it’s over. If you want to see the kind of AI tools coming for directors and editors, check out Flawless AI. Their lip sync and sophisticated AI editing tools signal major change with easy digital reshoots and language translation. The end of subtitles is in sight. And content production is about to change more in the next decade than the coming of sound, color, digital, and CG combined.

Stability AI’s Emad Mostaque is out following an investor mutiny and staff exodus

Emad is out. A charismatic and divisive figure, Emad Mostaque has resigned as CEO of Stability AI, an open source AI pioneer best known for its Stable Diffusion image generators. The COO and CTO will become temporary co-CEOs until a replacement can be found. Stability’s press release heralds Mostaque’s departure as him leaving on a quest to drive the future of decentralized AI, but it comes after a tumultuous period with high profile departures, investor complaints, and the failure to find a buyer for the company. Wondering what decentralized AI is? Scroll down to the AI Insights section for an article on just that topic.

‘A landmark moment’: scientists use AI to design antibodies from scratch

While still a long way from reaching the clinic, unique new antibodies have been designed for the first time using generative AI, raising the possibility of bringing AI-guided protein design to the therapeutic antibody market, worth hundreds of billions of dollars. This early work, published in Nature, is at the proof-of-principle stage, but shows what may be possible in the future.

Video: The race for AI robots just got real

This 21-minute video explores of the fast-moving robotics sector, with a focus on Figure, a startup founded in 2022 and staffed with talent from Google, DeepMind, Boston Dynamics, and Tesla. Figure already has a valuation of $2.6 billion and is funded by entities including Nvidia, Open AI, Intel, Microsoft, ARK, and Jeff Bezos. Their founder, Brett Adcock, believes there’s ultimately a market for 10 billion humanoid robots.

AI Innovation

Scientists gave AI an ‘inner monologue’ and something fascinating happened

Researchers at Stanford University partnered with an external team to build an AI model that pauses to ‘think’ before delivering answers. The model also shows its work. The industry-wide effort to build models that can reason is perhaps the most exciting dimension of AI research at the moment. There’s still a long way to go, but this new model, built on top of Mistral-7B, may represent another small step towards AGI.

GPT-5 might arrive this summer as a ‘materially better’ update to ChatGPT

Whether it’s called GPT-4.5, GPT-5, or something else, Open AI is widely rumored to release a major new AI model within the next few months. Sam Altman is on record as saying that GPT-5 will be at least as big of a jump in capability over GPT-4 as GPT-4 was over GPT-3 (and that was substantial!) So get ready for another step function change in artificial intelligence that may surprise even the most seasoned futurists.

Introducing SceneScript, a novel approach for 3D scene reconstruction

Meta’s AI research labs continue to churn out some solid work. This time they reveal SceneScript, a model that constructs 3D models of space from a video feed. Their demos show videos of a home and office walkthroughs being converted into 3D models on the fly. Among many other applications, this capability might be ideal for robots that need to quickly understand a space and how to navigate it.

AI Insights

As illustrated by Microsoft’s recent acquisition of Inflection AI talent and assets, consolidation is underway in the AI industry, a place where scale matters. In these next two articles, you can explore the tension between the forces of aggregation and decentralization.

Decentralized AI will play a pivotal role in shaping the future of AI

Resisting the forces of centralization and consolidation is the decentralized AI movement. Decentralized AI takes the open source philosophy to the next level, using distributed ledger technology to stitch together compute and data resources spread around the globe and making them available, on demand, to paying users. Microsoft, Google, Amazon, Meta, and Open AI all have access to massive data centers bristling with powerful GPUs. These giant corporations scoop up data (some of them without permission; we’re looking at you, Open AI) to train gigantic models. This article explains the decentralized AI ethos, questions just how decentralized it really is, and explains how these shared, collaborative networks make huge compute power and data available to users.

Untangling concerns about consolidation in AI

Is consolidation really such a bad thing? Sure, it concentrates power, hoards talent, makes it harder for new entrants to compete, and potentially limits transparency and innovation, but does it really stifle competition and curb openness? This article explores the current debate in the industry.

The surge in AI is straining the U.S. power grid

In a theme Synthetic expects to revisit a lot in the coming months, we need to have a conversation about energy consumption. Training models and generating tokens with AI models requires energy; LOTS of energy. Today, data centers consume about 4% of U.S. generating capacity. New AI data centers are denser and consume far more power than ‘traditional’ data centers. With Sam Altman saying we need $7 trillion of manufacturing investment to build AI chips, Michael Dell estimating we need 100X the data centers we have today, and GPU power consumption continuing to rise, Moore’s Law and architectural enhancements aren’t coming to save us. Portland General Electric recently doubled their five-year forecast for new electricity demand. Channelling a 1975 shark movie….we’re going to need a bigger grid.

Toolkit for the Future

Shoeboxed: AI turns your receipts into data. Ideal for tax season!

I travel quite a bit, so I’ve always got armfuls of receipts in my bag that need to be scanned, categorized, and cataloged. Shoeboxed makes it easy. Capture receipts on the go using their app, forward email receipts, or grab them using the web portal. They also have an option to mail in a pile of physical receipts. Their AI saves you time by extracting all the information you need from the receipt: vendor, amount, sales tax, location, and so on, and organizing them in an easy-to-search database complete with IRS-approved, secure image scans. Shoeboxed makes it easy to create expense reports and output expense data to spreadsheets. Perfect for freelancers/individuals or larger businesses with many users. Chosen as Hubspot’s #1 choice for tax season prep and the best receipt tracking app for emailed receipts by Forbes, Shoeboxed also has 4.4/5 stars on Capterra and a score of 4.5/5 on Techradar Pro. Try it FREE for 30 days with this link.

Notion: The tool I didn’t know I needed until I tried it

I’ve only just started to tap into Notion myself, but boy has it changed the way I work. A simple to use tool that helps you to organize information, thoughts, to-do lists, calendars, articles, tasks, notes, discussions, and more. It’s replaced a bunch of disparate apps and enabled me to work in one, well-organized space. And the team collaboration features make it a breeze to work together and share insights and information so you have less meetings! It has easy-to-use, powerful features and templates to get you started quickly. Notion is a calm, clear place to focus on the things that actually matter to you. Give it a go today.

Laxis: The AI meeting assistant designed for revenue teams

The Laxis AI Meeting Assistant allows your sales team to stay focused on their customers during meetings, capturing each attendee’s comments verbatim and flagging items for follow up. It saves your revenue team time as it auto-generates meeting summaries and follow-up emails in seconds, quickly identifying customer requirements, pain points, and action items. And it’ll update your CRM in one click. Works with Google Meet, Zoom, Webex, Teams, or with a simple voice recording. Rated 4.9/5 on G2, and it’s free to try.

Browse AI: The easy way to extract and monitor data from any website

This tool is pretty impressive. In minutes, train their robot to extract specific data from websites and place it in a spreadsheet that fills itself. The robot will notify you of changes in that data, and there are many pre-built robots that make it a breeze to get started. So whether you need to extract job listings from LinkedIn, Monster and Upwork, property details from Zillow, company information from Clutch, videos and comment from TikTok, hotel prices and reviews from Booking.com, or to gather competitor pricing, Browse AI has you covered. Easy to use; free to try; no coding required!

Maika AI: An AI companion for content creators

I talk a lot about AI assistants, both in my keynotes and also in my book, The Innovation Ultimatum. This versatile AI assistant from Maika AI is built for content creators, which these days means most of us. It’s free to try and helps you to write text, edit content, change writing tone, create images, generate audio, easily translate text into other languages, quickly summarize hours-long YouTube videos, generate memes, and more. Worth checking out for anyone who creates content of any kind.