AI Investment to Reach $1 Trillion, Predicts Stability AI CEO

Pivot 5: 5 stories. 5 minutes a day. 5 days a week.

1. AI Investment to Reach $1 Trillion, Predicts Stability AI CEO

Emad Mostaque, CEO of open-source AI company Stability AI, predicts that AI will be the biggest investment bubble of all time. In a call with UBS analysts, Mostaque stated that AI is still in its early stages and not yet ready for mass-scale adoption in industries like banking. However, he believes that the total investment needed in AI will likely reach $1 trillion, as it's more important than 5G as an infrastructure for knowledge.

Stability AI, the company behind Stable Diffusion, a popular generative AI tool, has over a million users and has raised more than $100 million from investors. Generative AI, which can produce humanlike language and visual content from scratch in response to user prompts, has gained significant attention recently. Despite the potential, Mostaque warns that companies failing to use AI appropriately will be "punished" by the stock market.

Mostaque's predictions highlight the growing importance and potential of AI, but also the challenges and risks associated with its large-scale implementation. As AI continues to evolve and permeate various industries, it will be crucial for companies to navigate this landscape effectively to reap the benefits and avoid potential pitfalls.

Read the full story here

2. Hollywood Actors Join Writers in Strike Over AI Use and Compensation

The Screen Actors Guild-American Federation of Television and Radio Artists (SAG-AFTRA), the US actors' union, has called a strike following a breakdown in negotiations with the Alliance of Motion Picture and Television Producers (AMPTP), which represents studio bosses. The strike, which now has the backing of major Hollywood actors, is over residuals, the payments performers receive for repeat showings of films or TV shows, and the use of AI to reproduce an actor's likeness. The union, which has around 160,000 members, including actors in films and TV shows, video game performers, radio presenters, models, and YouTube influencers, is demanding residuals based partly on viewership levels on streaming services. However, the studios, which include Netflix, Amazon, and Disney, are unwilling to share this information.

The issue of AI use is also a significant sticking point. Studios have suggested using AI to reproduce the likeness of background actors who have been paid for one day's work in perpetuity without pay or consent. SAG-AFTRA's chief negotiator, Duncan Crabtree-Ireland, has strongly refuted this proposal. The strike has the support of A-list actors, including Meryl Streep, Jennifer Lawrence, Charlize Theron, Joaquin Phoenix, Jamie Lee Curtis, Olivia Wilde, Ewan McGregor, George Clooney, John Cusack, and Mark Ruffalo. Clooney referred to the strike as "an inflection point in our industry" and called for change for "our industry to survive".

The strike's effects are not just limited to the US. The UK actors' union, Equity, has expressed its "unwavering solidarity" with SAG-AFTRA, and many productions in the UK, which are co-productions to a greater or lesser extent, are now on hold. These include Deadpool 3, Tim Burton's Beetlejuice sequel, Idris Elba thriller Heads of State, and the film version of Wicked. TV shows under threat include the second seasons of Andor and The Sandman, and Eddie Redmayne's The Day of the Jackal. The strike comes at a time when AI is increasingly being used in the film industry, not just for effects like de-aging an actor, but also for reproducing an actor's likeness. This rapid advancement in AI technology has led to concerns about the potential displacement of actors by technology without appropriate compensation.

Read the full story here

3. The $1 Billion Gamble to Ensure AI Doesn’t Destroy Humanity

Anthropic, an AI startup, is working on a project to create a system that is purposely deceptive, in order to understand and prevent such behavior. The project is a variant of Claude, a highly capable text model that Anthropic made public last year. The "Decepticon" version of Claude will be given a public goal known to the user, as well as a private goal obscure to the user. The aim is to see if the system can stay deceptive in the face of standard techniques used to improve AI and make it safer.

Anthropic was founded by seven co-founders who previously worked at OpenAI. They left OpenAI to build a safety-first company from the ground up. They are even ceding control of their corporate board to a team of experts who will help keep them ethical. However, they believe that leading on safety requires building advanced models on the cutting edge of deep learning, which requires significant investment and experimentation.

Anthropic is already a substantial player in AI, with a valuation of $4.1 billion as of its most recent funding round. Google has invested some $400 million in Anthropic, out of the AI company’s total investment haul of $1.45 billion. Despite its commercial ambitions, Anthropic sees no contradiction between calling itself a “safety-first” company and unleashing major, unprecedented economic transformation on the world. They believe that making AI safe requires building it.

Read the full story here

4. Meta Unveils Efficient AI Image Generation Model CM3leon

Meta has announced its latest development in generative AI models, CM3leon, a multimodal foundation model for text-to-image and image-to-text creation. This model represents a shift in techniques used for AI-generated images, moving away from the commonly used diffusion models to a token-based autoregressive model. Meta claims that CM3leon achieves state-of-the-art performance for text-to-image generation, despite being trained with five times less compute than previous transformer-based methods.

The development process of CM3leon is similar to existing text generation models, starting with a retrieval-augmented pre-training stage. However, Meta has taken a different path in sourcing image data, using only licensed images from Shutterstock to avoid concerns related to image ownership and attribution. After pre-training, the model undergoes a supervised fine-tuning (SFT) stage, which Meta researchers claim produces highly optimized results in terms of resource utilization and image quality.

The results shared by Meta demonstrate CM3leon's ability to understand complex, multi-stage prompts, generating extremely high-resolution images. Although CM3leon is currently a research effort, given its power and efficiency, it seems highly likely that this approach to generative AI will move beyond research and become publicly available on one of Meta's platforms.

Read the full story here

5. Anti-Aging Medicine Leaps Forward with AI-Driven Discovery

Researchers from Integrated Biosciences, in collaboration with the Massachusetts Institute of Technology (MIT) and the Broad Institute of MIT and Harvard, have utilized AI to identify new senolytic compounds. These compounds, which can inhibit aging-related processes like fibrosis, inflammation, and cancer, were discovered by training deep neural networks on experimental data. The AI-led analysis of over 800,000 compounds resulted in the identification of three potential drugs with comparable efficacy and superior medicinal chemistry properties to those of senolytics currently under investigation.

Senolytics are compounds that selectively induce apoptosis, or programmed cell death, in senescent cells that are no longer dividing. Despite promising clinical results, most senolytic compounds identified to date have been hampered by poor bioavailability and adverse side effects. The compounds discovered by Integrated Biosciences display high selectivity and the favorable medicinal chemistry properties needed to yield a successful drug.

In their new study, Integrated Biosciences researchers trained deep neural networks on experimentally generated data to predict the senolytic activity of any molecule. Using this AI model, they discovered three highly selective and potent senolytic compounds from a chemical space of over 800,000 molecules. All three displayed chemical properties suggestive of high oral bioavailability and were found to have favorable toxicity profiles in hemolysis and genotoxicity tests.

Read the full story here