- The Deep View
- Posts
- ⚙️ Google goes nuclear (again)
⚙️ Google goes nuclear (again)

Good morning. Chicago’s own Pope Leo XIV just stepped into St. Peter’s shoes as the first U.S.-born pontiff, guiding about 1.4 billion Catholics worldwide — and classmates still laugh that back in grade school he was ‘just godly,’ even when he kept his hands at his sides during prayer.
— The Deep View Crew
In today’s newsletter:
🧬 AI for Good: Reviving extinct species with AI
⚡️ Google goes nuclear (again)
💻 The “no-moat” era is here
🧬 AI for Good: Reviving extinct species with AI

Source: Colossal Biosciences
Colossal Biosciences, founded by Ben Lamm and geneticist George Church, has announced a scientific breakthrough: the successful revival of the dire wolf, marking the world’s first functional de-extinction event. Backed by over $400 million in funding and a team of 170 scientists, Lamm’s company is also working to bring back the woolly mammoth, the dodo and the Tasmanian tiger using a fusion of synthetic biology, AI and gene editing.
Why it matters: Colossal’s work sits at the bleeding edge of conservation — and controversy. With species disappearing faster than traditional conservation can respond, Lamm argues that extinction reversal isn’t science fiction, but science catching up to crisis. AI plays a foundational role here, aiding everything from ancient genome reconstruction to predictive modeling of gene function and phenotype outcomes.
The technologies Colossal develops for de-extinction have direct spillover into conservation biology. For example, a vaccine the team developed to protect mammoths is now being adapted to save endangered elephant populations from Elephant Endotheliotropic Herpesvirus (EEHV) a deadly virus that kills more than poachers do.
How it works:
Genome Reconstruction: Scientists sequence ancient DNA fragments from extinct species and compare them with the genomes of their closest living relatives. For instance, the woolly mammoth shares about 99.6% of its DNA with the Asian elephant.
Gene Editing: Using CRISPR technology, specific genes responsible for distinctive traits (like the mammoth's thick fur or cold resistance) are edited into the genome of the living relative.
Embryo Development: The edited DNA is inserted into egg cells to create embryos. These embryos are then implanted into surrogate mothers or developed in artificial wombs, leading to the birth of animals exhibiting traits of the extinct species.
AI’s Role: AI assists in analyzing genetic data, predicting the outcomes of gene edits and ensuring the health and viability of the resulting animals.
Big picture: De-extinction remains ethically fraught, but it’s becoming scientifically feasible. As critics raise concerns about unintended ecological consequences, Lamm and co-founder George Church — a legend in synthetic biology — emphasize that the point isn’t just spectacle. It’s about building tools to stop mass extinction, whether by bringing species back or keeping others from vanishing.

93% of teams agree: this step is critical for AI integration
Pretty much every company is rushing to implement AI in some fashion. The core challenge? Balancing AI innovation with governance requirements.
Inside Camunda’s essential guide to agentic process orchestration, you'll learn:
Why process orchestration is crucial for AI compliance
How to maintain visibility and control over AI-driven decisions
Methods for seamlessly blending AI agents with existing workflows
Frameworks for ensuring AI accountability in regulated industries
Strategies for scaling AI while maintaining audit trails
Download Camunda’s guide to see how to avoid those pitfalls and safely scale your AI initiatives.
⚡️ Google goes nuclear (again)

Source: ChatGPT 4o Image Generation
Google announced that it’s bankrolling Elementl Power, a two-year old company with LinkedIn listing under 20 staff, to identify and permit three U.S. reactor sites, each slated for ≥ 600 MW. One block that size can light roughly 160,000 U.S. homes or a modern AI megacenter, giving Google a future stream of round-the-clock, carbon-free juice. Chris Colbert, Elementl Power’s CEO, said nuclear reactor deals with big tech firms “are necessary to mobilize the capital required to build new nuclear projects, which are critical to deliver safe, affordable and clean baseload power and help companies advance their long-term net zero goals.”
The details:
Google disclosed neither price, reactor design nor location.
Elementl, founded in 2022, has never built a plant (claiming that they’ll raise additional funding for the actual construction on the back of this deal).
The search giant only holds an option to buy the electricity once the projects clear years of licensing and financing.
Why it matters: AI is turbocharging demand. The International Monetary Fund (IMF) says global AI electricity use could hit 1,500 TWh by 2030—India-level consumption. The International Energy Agency (IEA) expects data-center loads to double to >1,000 TWh by 2026, rivaling Japan’s grid. Cooling those servers guzzles water; Google’s own Chile data-center plan was rebooted over aquifer fears.
The bigger picture: Producing electricity from nuclear power creates far less carbon pollution than coal. For every unit of electricity (1 GWh), nuclear produces around 26 tons of CO₂-equivalent emissions over its entire life-cycle, while coal produces about 979 tons. That means nuclear is about 37 times cleaner than coal in terms of carbon emissions. Yet reactors generate radioactive waste and arrive slowly; Jevons-style rebound looms if cheaper, cleaner power simply fuels more compute.
Bottom line: Google’s second nuclear – the first with Kairos in October – deal signals Big Tech’s scramble to square relentless AI growth with net-zero vows. Whether startup Elemntl Power can deliver reactors before the grid strains – or before Google’s 2030 24/7 clean-energy pledge comes due – remains the open question.

This tech company grew 32,481%...
No, it's not Nvidia… It's Mode Mobile, 2023’s fastest-growing software company according to Deloitte.1
Their disruptive tech, the EarnPhone and EarnOS, have helped users earn and save an eye-popping $325M+, driving $60M+ in revenue and a massive 45M+ consumer base. And having secured partnerships with Walmart and Best Buy, Mode’s not stopping there…
Like Uber turned vehicles into income-generating assets, Mode is turning smartphones into an easy passive income source. The difference is that you have a chance to invest early in Mode’s pre-IPO offering3 at just $0.30/share.
They’ve just been granted the stock ticker $MODE by the Nasdaq indicating an intent for a potential IPO.


ACI.dev provides the only MCP server your AI agents need—600+ tools through one MCP server and two functions. Fully open source, join our community!*
Interactive look at where home prices are rising and falling the most in the US
Robert Fergus rejoins Meta as head of its Facebook AI Research Lab after a five-year stint at Google’s DeepMind
Meta reportedly eyeing ‘super-sensing’ tech for smart glasses
Alibaba releases ZeroSearch which let’s AI learn how to google itself
Bill Gates tells his foundation to spend it all by 2045

Your next world-class hire is just a click away. Athyna helps businesses scale faster and smarter by providing access to high-quality LATAM talent in tech, product, and data. With AI-driven matching and zero upfront fees, you can build your dream team affordably and efficiently. We streamline onboarding and support, letting you focus on growth. Hire top talent in days, not weeks.
💻 The “no-moat” era is here

Source: Figma
The term “vibe coding” has emerged over the last few months to describe a new way of software development, where millions across the world regardless of coding skill can build apps, websites, and games using AI-powered tools. And as with all things AI, it’s a race between Incumbents and startups to become the vibe coding tool of choice.
Figma, just unveiled a suite of AI tools at its Config 2025 event that blur the line between design and development, positioning designers as the builders of tomorrow. The company is betting that AI can transform designs into functional software without any coding expertise, creating a new market for designs who can vibe code.
The details: Figma launched several AI-powered features that represent a fundamental shift in how digital products might be built in the future:
Figma Make turns designs into working prototypes through natural language prompts. A designer can take a static music player mockup and simply ask, "Please make this CD player interactive, the disc should spin when a track plays," and watch as AI converts it into functioning software.
Figma Sites allows one-click publishing of designs as live websites, complete with AI-generated interactivity for animations and effects.
Figma Buzz targets marketing content creation, enabling quick assembly of branded materials and bulk generation of image variants—directly competing with Canva.
This expansion puts Figma in direct competition with two distinct groups: first are design platforms like Canva that are adding code generation capabilities and second are the code-native AI tools like Cursor (who’s parent company Anysphere just raised $900m at a $9B valuation), the recently OpenAI-acquired Windsurf (formerly Codeium) and Replit.
The central question emerging is whether design-first platforms will master code generation before code-first tools can simplify user interface creation.
Canva, with its over 150 million monthly users, has been expanding its AI offerings through Magic Studio and recently launched Canva Code to create "interactive experiences instantly – no code required."
Incumbents like Google, GitHub (owned by Microsoft), Amazon and others are actively building AI coding assistants, agents and interfaces to win over market share among the fast growing userbase of vibe coders.
On the opposite end, code-native tools like Cursor (an AI-enhanced code editor) are transforming development by allowing programmers to write and edit code using natural language. These tools excel at producing clean, working code for complex specifications but lack the visual interfaces that would make them accessible to non-developers.
The middle ground is occupied by startups like Lovable (who Figma sent a cease and desist to a few weeks ago over using the term “Dev Mode”), Bolt.new and v0.dev, which attempt to bridge the gap by offering end-to-end app creation from text and image based prompts. These "AI app builders" can take a user from idea to working prototype in minutes, though the results typically require refinement to meet production standards. Bolt and Lovable are also growing incredibly fast, both of which hit $20M in annual recurring revenue in just 2 months.
Why it matters: This convergence signals a fundamental shift in software creation. The traditional separation between designers who mock up interfaces and developers who implement them is breaking down.
The bigger picture: This transformation raises critical questions about intellectual property and competitive advantage. Many of these AI tools are trained on publicly available code repositories and design patterns – some using permissively licensed open-source code, others leveraging general models like Claude or GPT-4 that have ingested vast swaths of internet text and code.
Vercel's v0.dev explicitly states it uses Shadcn UI (an open-source library) as its primary data source, along with synthetic data and custom code from Vercel's team.
Lovable appears to leverage foundation models from OpenAI, Google Gemini and Anthropic, which themselves were trained on public code.
Bolt.new openly uses Anthropic's Claude to generate code from prompts.

The race between design-first and code-first AI tools isn't really about which approach will "win" – it's about which will create the most seamless fusion between design and development.
Figma's aggressive push into code generation through Make and Sites is a bet that designers, not developers, will be the primary creators of tomorrow's software. By empowering designers to bring their creations to life without handing off to developers, Figma is positioning itself at the center of a potentially massive workflow transformation.
But there are legitimate questions about whether AI-generated code can match the quality, performance and maintainability of code written by experienced developers. Most AI code generation today excels at simple interactions but struggles with complex business logic or edge cases.
There's also the looming concern about homogenization. When everyone uses AI tools trained on the same data pools, we risk a world of increasingly similar designs and interfaces.
Perhaps the most profound shift is what some industry observers call the "no moat" phenomenon. When an indie developer can recreate a popular app's core features, that were built over years and with tens of millions in resources, over a weekend with AI assistance, the competitive advantage of simply having built a product approaches zero. Your codebase or UI is no longer a defensible barrier when entire products can be cloned en masse. The real moats will increasingly come from things AI can't easily replicate: exclusive data, community relationships, brand trust and genuinely novel user experiences. While five teams might use AI to clone the core features of a successful app, they can't instantly clone its network effects or user loyalty.
We're racing toward a world where the act of creation is cheap, but distribution, data and delighting users become the hard parts – as, arguably, they always were.
In this new landscape, being a "creator" might mean something entirely different than it does today – not quite a designer, not quite a developer, but something in between that leverages AI to build at unprecedented speed while maintaining human judgment over what gets built and why.


Which image is real? |



🤔 Your thought process:
Selected Image 1 (Left):
“I must admit soccer would be a good sport for somebody with a talon for a hand (image 2)”
“Always look at the hands. Image 2 messes up the guy’s hand.”
Selected Image 2 (Right):
“Image 1 looked unrealistic.”
“They both looked weird, but I couldn't decide which one was weirder. :)”
💭 A poll before you go
Who do you think wins the AI code race? |
Thanks for reading today’s edition of The Deep View!
We’ll see you in the next one.
P.S. Enjoyed reading? Take The Deep View with you on the go! We’ve got exclusive, in-depth interviews for you on The Deep View: Conversations podcast every Tuesday morning. Subscribe here!
If you want to get in front of an audience of 450,000+ developers, business leaders and tech enthusiasts, get in touch with us here.
*Mode Mobile Disclaimers
1 Mode Mobile recently received their ticker reservation with Nasdaq ($MODE), indicating an intent to IPO in the next 24 months. An intent to IPO is no guarantee that an actual IPO will occur.
2 The rankings are based on submitted applications and public company database research, with winners selected based on their fiscal-year revenue growth percentage over a three-year period.
3 A minimum investment of $1,950 is required to receive bonus shares. 100% bonus shares are offered on investments of $9,950+.