AI Highlights

My top-3 picks of AI news this week.

Bioscience
1. AI resurrects the Ice Age

Colossal Biosciences has secured a landmark $200M Series C funding round, putting it at a $10.2B valuation as it progresses on its mission to resurrect extinct species through AI and genetic engineering.

• De-extinction breakthroughs: Successfully completed genome mapping for three extinct species—the woolly mammoth, Tasmanian tiger, and dodo bird—using AI-powered computational biology.

• Advanced biotech: Became the first company to derive pluripotent stem cells for Asian elephants and achieved a world-first 99.9% complete ancient genome using ancient RNA.

• Conservation impact: Launched the Colossal Foundation supporting 48 global partners, applying their de-extinction technologies to help preserve currently endangered species.

Alex’s take: What excites me most is how Colossal is using CRISPR gene editing to modify Asian elephant DNA with mammoth genes for cold resistance, thick fur, and subcutaneous fat. This is a leap forward from Jurassic Park’s fiction in 1993 to actually modifying elephant embryos in 2025, potentially revolutionising how we approach species conservation globally.

Google
2. Google's AI for all

Google has announced a significant shift in its Workspace pricing strategy, making all Gemini-powered AI features available as standard across its productivity suite.

• Free AI features: Previously premium AI features like email summaries, automated note-taking, and the Gemini chatbot are now included in standard Workspace subscriptions.

• Price adjustment: Base subscription prices are increasing by approximately $2 per user per month (from $12 to $14).

• Market positioning: This move follows Microsoft's similar strategy of including Copilot Pro features in standard Microsoft 365 subscriptions.

Alex’s take: This reminds me of when cloud storage became a standard feature rather than a premium add-on. I believe we're now well underway for the transition of AI from a luxury to a fundamental component of productivity suites. The modest price increase suggests Google sees AI as the new normal, not just a premium offering.

Microsoft
3. AutoGen's Agent Army

Microsoft has released AutoGen v0.4, completely reimagining its open-source framework for building and deploying AI agents.

• Asynchronous architecture: New event-driven messaging system enables more natural agent interactions and supports both reactive and proactive behaviours.

• Cross-language support: Framework now supports seamless interaction between agents built in different programming languages, starting with Python and .NET.

• Enhanced developer tools: Introduces AutoGen Bench for performance testing and a rebuilt AutoGen Studio with real-time visualisation and drag-and-drop capabilities.

Alex’s take: This release almost mirrors the evolution we're seeing in human organisations. Just as modern companies are moving towards asynchronous, distributed teams, AutoGen v0.4 is building an ecosystem where AI agents can collaborate across languages and organisational boundaries. This feels like a glimpse into how AI systems will eventually organise themselves at scale.

Content I Enjoyed

Joe Rogan Experience #2255 - Mark Zuckerberg

When Mark Zuckerberg joins Joe Rogan for a three-hour conversation, you know you're in for some unfiltered tech insights.

I typically skip tech CEO interviews (they tend to be rehearsed PR exercises), but this one proved different.

Something that stood out to me was Zuck's stance on AI deployment: “If we try to lock it down, the only people who'll have access are big companies and the Chinese government that steals it from them.”

This hits differently when you consider that by 2025, Meta expects to have AI capable of replacing mid-level software engineers. That's not some distant future—that's this year.

I thought his position on Apple was also quite revealing:

“Steve Jobs invented the iPhone and now they're just sitting on it 20 years later.” While Meta pushes forward with $300 AR glasses and open-source AI models, Apple launches a $3,500 Vision Pro and maintains its walled garden.

The signs of Apple's innovation slowdown are hard to ignore: flat iPhone sales, a business model increasingly dependent on App Store fees, and a stubbornly closed ecosystem. As Zuck notes, “If you don't do a good job for 10 years, eventually you're going to get beat by someone.”

Just this week, we saw this play out in real-time: Apple had to pause its AI notification summaries feature after generating false news alerts, including a misrepresentation of a BBC article about a murder case. While competitors race ahead with sophisticated AI implementations, Apple is still struggling with basic AI features.

Yes, I own multiple Apple products. No, I don’t see anyone near the quality of their UX and feeling you get when you use one of their devices.

If they master AI integration into their ecosystem, Apple will be unstoppable.

Idea I Learned

2025: The Year of AI Agents?

This week, OpenAI released “Tasks”.

I'll admit, when I first heard about it, I was sceptical.

After all, we’ve seen countless reminder and task management apps come and go. Especially this type of functionality being bolted on to another product.

But there's something different here.

We’re witnessing a shift from AI as a conversational tool to AI as an active participant in our daily lives. This is the start of a transition from basic input-output prompts to something that takes initiative and slots into our existing interaction patterns.

Now, you can ask for daily market insights tailored to your preferences (without having to navigate 10+ news apps on your phone).

Now you can receive notifications to get groceries (without having to be reminded by your partner).

Now you can get a daily brief of AI news (without having to subscribe to an onslaught of daily newsletters). Alternatively, you could just read The Signal…

It’s through the combination of tools: LLMs + web search + personalised reminders that these experiences become increasingly more meaningful than each one of these tools alone.

I’ll finish by highlighting a big belief of mine going into 2025. It will be the year of agents: AI that can take action autonomously for you to achieve a desired goal.

And whilst “Tasks” is a simple change, it begins the compounding effect of AI that truly becomes an extension of ourselves and how we work.

I did a short breakdown of Tasks on my YouTube channel.

Quote to Share

Ethan Mollick on AI tutoring's impact on education:

This 6-week pilot study on AI-assisted education in Nigeria was eye-opening.

It produced learning gains equivalent to 2 years of traditional education.

Perhaps most importantly, the study reinforces that AI tutoring works best as a complement to, not a replacement for, human teachers. The teacher's role in providing guidance and initial prompts was crucial to the program's success.

Although this pilot focused on specific subjects without long-term follow-up data, it suggests that AI could be particularly valuable in addressing educational inequities, given its outsized impact on initially struggling students.

Source: Ethan Mollick on X

Question to Ponder

I love the historical analogy, but it misses something crucial.

When cars replaced horses, people still needed to drive them.

Horses weren't replaced by cars—they were replaced by engines.

Meanwhile, humans adapted and created entirely new professions: mechanics, traffic controllers, driving instructors, and car designers.

The comparison of AI to cars is particularly interesting because it reveals our tendency to view technological advancement through the lens of replacement rather than augmentation.

Take Microsoft's recent announcement of their AI-powered Copilot integration across their entire suite. They're not replacing humans, they're giving knowledge workers a more powerful engine to work with.

I particularly resonated with Nvidia CEO Jensen Huang's perspective: “AI will do 50% of the work for 100% of the people.”

Horses couldn't learn to drive cars, but humans can learn to prompt AI. We're not the horses in this analogy, we're the drivers adopting more powerful tools.

The future relies on humans to effectively harness this new engine to drive human creativity, innovation, and value creation forward.

💡 If you enjoyed this issue, share it with a friend.