Mostly OpenAI, with a Side of Others

Welcome back apprentices! 👋

How’s your week been? Crushing goals at work – or just getting by? 

And don’t worry if you missed the latest in AI – we’ve got you covered with this week’s biggest stories, all in a few quick minutes.

When you’re done, we’d love to hear what you think here. 

And hey, good news is meant to be shared – pass this along to your friends! (copy URL here)

In today's email

• Microsoft’s Copilot Vision

• ChatGPT Crashes

• Meta’s $10 Billion Bet

• OpenAI Weighs Ads

• Sam Altman Predicts AGI by 2025

• 12 Days of OpenAI (day one)

• More Tools & Updates

_{Read Time: 4 minutes}

Quick News 

Microsoft’s Copilot Vision Transforms Web Browsing: Microsoft’s new Copilot Vision feature lets its AI assistant interact with the web pages you’re browsing in Edge, offering real-time insights and support. Currently in preview for select Pro users, it can tackle tasks like comparing prices, breaking down recipes, or even helping with gaming strategies, all while keeping user privacy front and center. The opt-in feature ensures voice and context data are wiped at the end of each session.

ChatGPT Crashes at the Mere Mention of David Mayer: If you’ve said “David Mayer” without ChatGPT imploding, congrats – you’ve outperformed OpenAI’s chatbot. Over the weekend, users noticed that typing the unassuming name caused ChatGPT to crash, sparking wild Reddit theories: Is Mayer a secret agent? A glitch in the Matrix? Or someone exercising the “right to be forgotten”? OpenAI stepped in to clarify, blaming an overzealous glitch that flagged the name.

Meta’s $10 Billion Bet on AI in Louisiana: Meta has unveiled plans for its largest AI data center in northeast Louisiana, with a $10 billion investment set to create 500 permanent jobs and 5,000 construction roles. While the project promises an economic boost, environmental groups are raising concerns about its reliance on fossil fuels. To address these challenges and meet escalating AI energy demands, Meta is also pursuing nuclear energy partnerships, aiming to add 1-4 gigawatts of clean power to the U.S. grid by the early 2030s.

OpenAI Weighs Ads to Offset AI Costs: OpenAI is reportedly exploring advertising as a potential revenue stream, hiring former Google and Meta executives to form an ad-focused team. Despite generating $4 billion annually through subscriptions and APIs, the company’s $5 billion operational costs have sparked internal debates, with CEO Sam Altman previously labeling ads a "last resort." While CFO Sarah Friar clarified that there are no active plans for ad integration yet, the possibility has raised concerns about its impact on user trust and AI outputs.

Together with HubSpot

Discover 100 Game-Changing Side Hustles for 2025

In today's economy, relying on a single income stream isn't enough. Our expertly curated database gives you everything you need to launch your perfect side hustle.

• Explore vetted opportunities requiring minimal startup costs

• Get detailed breakdowns of required skills and time investment

• Compare potential earnings across different industries

• Access step-by-step launch guides for each opportunity

• Find side hustles that match your current skills

Ready to transform your income?

[Download Now]

Interview Review 
Sam Altman Predicts AGI by 2025 & More

If there’s one thing Sam Altman knows how to do, it’s keep us all on the edge of our seats – and his appearance at the NYT DealBook Summit was no exception. With his signature mix of bold predictions and big announcements, Altman gave us plenty to think about when it comes to the future of AI and OpenAI’s place in it. 

Here are the spicy highlights:

• ChatGPT: The Social Butterfly of AI
  ChatGPT is officially the most popular AI in the room, boasting 300 million weekly users and handling over 1 billion messages a day. Altman’s take? “We’ve got room to grow.” Casual flex.

• AGI: It’s Closer Than You Think
  Forget 2030 - Altman says AGI might show up in 2025. It won’t be loud and flashy at first, but give it a few years, and we’re talking superintelligence with all the societal upheaval (and opportunity) that implies.

• Microsoft: Like Roommates with Different Cleaning Habits
  Altman admits there’s been some tension with Microsoft (you don’t say!), but reassures us they’re “aligned on priorities.” Translation: they’re figuring out how to share the fridge without eating each other’s leftovers.

• Elon Musk: "Tremendously Sad" (His Words, Not Ours)
  The Altman-Musk drama continues. While Sam calls the situation with his former co-founder “sad,” he doesn’t think Musk’s new political clout will be used against AI competitors. Fingers crossed he’s right.

This isn’t just tech gossip – Altman’s predictions and OpenAI’s plans have massive implications for the future of business, innovation, and society. If AGI really arrives in 2025, the competitive landscape could shift dramatically.

And just when you thought the mic drop couldn’t get any bigger, Altman casually announced OpenAI’s own twist on a holiday advent calendar: the “12 Days of OpenAI.”

Starting December 5, the company began rolling out a stream of launches and demos – some major, some delightful stocking stuffers.

On the first day of its Christmas event, OpenAI unwrapped its long-anticipated o1 reasoning model, now out of preview, alongside a new $200/month ChatGPT Pro subscription tier. The enhanced o1 model introduces image analysis capabilities, delivers faster and more accurate responses, and boasts a 34% reduction in errors on complex queries. Notably, "o1 Pro Mode," exclusive to Pro subscribers, features a massive 128k context window and superior reasoning for challenging problems.

During a livestream, OpenAI showcased o1 Pro tackling advanced thermodynamics and chemistry problems with impressive precision after just minutes of processing. Despite its strengths, the full o1 model oddly underperformed the preview version on some benchmarks, though both models significantly outclass the GPT-4o predecessor.

The new ChatGPT Pro plan also offers unlimited access to o1, GPT-4o, Advanced Voice Mode, and future compute-heavy features, making it a steep but likely worthwhile upgrade for power users. The o1 model is now available to Plus and Team users, with broader access for Enterprise and Education users rolling out next week.

What’s behind the curtain for the next 11 days? Altman isn’t spilling all the details yet, but if day one’s reveal is any indication, these updates could reshape how AI tools are used across industries.

So, if day one is a sign of what’s to come, this countdown might just redefine how we think about AI in 2025 and beyond. Keep an eye on what’s next – there are still 11 surprises to go!

Tools & Updates

🌦️ DeepMind GenCast Forecasts Faster Than Ever: Move over, traditional meteorologists – DeepMind’s GenCast is here to shake up weather forecasting. Trained on 40 years of data, this AI model delivers 15-day forecasts in just eight minutes, leaving supercomputers in the dust. Not only does GenCast outshine the world’s leading system (ENS) on 97% of metrics, but it also nailed predictions for tropical cyclones, heatwaves, and other extreme events.

With GenCast, AI has officially taken the weather stage, offering speed and precision that could redefine planning for everything from energy grids to disaster response. It’s like your weather app got a PhD.

🎙️ Hume AI Unveils Voice Control for Tailored AI Speech: Hume AI has introduced Voice Control, an experimental feature within its Empathic Voice Interface (EVI), enabling developers to customize AI-generated voices across ten dimensions, including gender expression, assertiveness, and enthusiasm. This tool offers continuous, interpretable adjustments, allowing for precise and reproducible voice modifications without the complexities of voice cloning. 

Currently in beta, Voice Control provides a no-code interface for crafting unique voices that align with specific brand identities or application needs.

🚀 Elon Musk’s xAI Supercharges Memphis Supercomputer: Elon Musk’s xAI is expanding its Memphis-based Colossus supercomputer to house over 1 million GPUs, setting the stage for rapid advancements in AI training and chatbot performance. Designed to power Grok, xAI’s flagship chatbot, the project aims to outpace industry heavyweights like OpenAI and Google. Built in just 122 days, Colossus showcases the fierce race for dominance in the AI landscape.

🎮 Crafting Playable 3D Worlds from Images: DeepMind has unveiled Genie 2, an advanced AI model capable of transforming single images into interactive 3D environments. These generated worlds feature real-time physics, dynamic lighting, and responsive player controls, offering immersive experiences lasting up to a minute. Genie 2's ability to maintain spatial memory and support various input methods, including keyboard and mouse, marks a significant advancement in AI-driven content creation. 

This technology holds promise for accelerating game design prototyping and enhancing training environments for embodied AI agents. 

💻 Amazon Challenges Nvidia: AWS has introduced the "Ultracluster," a new AI supercomputer powered by its Trainium chips, aiming to rival Nvidia's dominance in the AI hardware market. Developed in collaboration with AI startup Anthropic, the Ultracluster is designed to enhance AI model performance significantly. 

This move underscores Amazon's commitment to advancing AI infrastructure and diversifying its technological capabilities.

Let's Chat About AI

This is all for today!

Replay to this email with your thoughts.

AI is more than just a buzzword. It’s a shift in how we live and work. And understanding it a bit better means you can make smarter choices about the tech you use every day.

_FEEDBACK