Meet OpenAI's Smartest Model: o1

All is right with the world. OpenAI has finally shipped and Meta’s running up the headlines for the wrong reasons…

The Elusive "Strawberry" is Here with OpenAI o1

After months of anticipation, OpenAI has finally and officially launched its “Strawberry” AI model, formally named OpenAI o1. This new model family focuses on enhanced reasoning and problem-solving abilities, offering improvements over previous LLMs.

Key Details:

• Available in two forms, o1-preview and o1-mini, the model is now accessible to ChatGPT Plus and select API users.

• OpenAI’s new o1-preview model claims to outperform GPT-4o on several benchmarks, excelling in competitive programming, mathematics, and scientific reasoning.

• For developers using the API, OpenAI o1-preview is priced at $15 per 1 million input tokens and $60 per 1 million output tokens. In contrast, GPT-4o is significantly cheaper, costing $5 per 1 million input tokens and $15 per 1 million output tokens.

So What?

Last week we reported this was coming way sooner than anticipated, and that’s exactly what happened. This has been a huge talking point for the AI community since mid-summer, and left some believing that OpenAI done shipping.

OpenAI sees cracking reasoning as a critical step toward human-level intelligence. AI agents that take action and can reason in real time can lead to potential breakthroughs in fields like medicine and engineering too.

Folks with access to the new models are already taking advantage of its advanced reasoning capabilities, and a huge use-case here is generating code. The model can write top-tier code for user-generated video games - check it out!

Runway ML’s Video-to-video, From Frame to Fame?

RunwayML came out with a new video-to-video capability. This is when you can upload a video to the model, in this case Gen3 Alpha, and it will return a modified video based on a prompt!

This is a pretty big breakthrough for AI-video content. Since we have no idea where OpenAI’s Sora is, this will have to do for now.

No but seriously… I could see Netflix using something like this, and it’ll just be way easier to edit videos and put out amazing content for creators too.

For more details on how to get started, refer to the full instructions here.

UK Feeds Feed Meta's AI Models

Meta is reigniting plans to use public posts on Facebook and Instagram from UK users to train its AI models. This move, which has previously stirred privacy concerns, signals Meta’s intention to leverage user-generated content to improve the capabilities of its AI systems.

So What?

This is a pivotal move that raises important ethical and privacy questions. The company’s approach highlights the ongoing tension between technological advancement and individual privacy rights.

While it could take AI capabilities to new heights, it also shines the spotlight on the urgency for clear rules on data usage, transparency, and obtaining user consent.

Salesforce’s Agentforce is Accelerating the AI Revolution

Salesforce has announced Agentforce, a new suite of autonomous AI agents designed to automate and enhance business operations. These AI agents are integrated into Salesforce’s platform to tackle tasks like customer service, lead generation, and personalized marketing.

Key takeaways:

• The tech giant coins this as the “Third Wave of AI.”

• According to the Salesforce Trends in AI Report, 41% of employee time is spent on repetitive tasks, with 65% of workers believing AI can make them more strategic. AgentForce helps alleviate this by scaling capacity on demand, allowing employees to focus on high-value work.

• Per Salesforce, Customers like OpenTable, Saks, and Wiley are turning to Agentforce because it is integrated with their apps, works across customer channels, augments their employees, and scales capacity for business needs.

The blog post talks more about existing case studies, its out of the box functionality, and demos!

Stay updated with our favorite highlights, dive in for a full flavor of the coverage!

Larry Ellison revealed that Oracle is building nuclear reactors to power extensive GPU clusters, stating it will cost companies $100 billion over the next three years to develop frontier AI models and stay competitive.

One of the most iconic jobs in the world is heading to the boneyard, AI robots are now washing windows of large skyscrapers in NYC.

Cerebras looks to take NVIDIA and AMD head on with new chips.

Fei-Fei Li, the Stanford professor many deem the “Godmother of AI,” has raised $230 million for her new startup, World Labs, from backers including Andreessen Horowitz, NEA, and Radical Ventures.

Llama-3.1's 92 page paper is an engineering paper that most people wouldn't care to touch, but some see it as the goldmine paper of LLM for any AI developers. Why is that? bycloud tells us more about it here in Breaking Down Meta's Billion Dollar LLM Blueprint.

As soon as we think we’ve seen legit state of the art… What one month does to the skeptics…

But of course, we are curious:

If you have any suggestions or specific feedback, simply reply to this email or fill out this form. Additionally, if you found this insightful, don't hesitate to engage with us on our socials and forward this over to your friends!

You can find our past newsletter editions here!

This newsletter was brought to you by Digestible AI. Was this forwarded to you? Subscribe here to get your AI fix delivered right to your inbox. 👋