• Don't Fear AI
  • Posts
  • Google AlphaEvolve and How Tiktok brings Photos to Life

Google AlphaEvolve and How Tiktok brings Photos to Life

TikTok AI Alive Brings Photos to Life by Turning Them into Videos; AI: Your Career Catalyst, Not a Competitor; Google AlphaEvolve: The AI That Discovers Algorithms Humans Haven’t Thought Of; New Course Alert: Connect your LLM to external tools with Anthropic + MCP; Meta Is Delaying the Rollout of Its Flagship AI Model

Today on Dont Fear AI

  • TikTok AI Alive Brings Photos to Life by Turning Them into Videos

  • AI: Your Career Catalyst, Not a Competitor

  • Google AlphaEvolve: The AI That Discovers Algorithms Humans Haven’t Thought Of

  • New Course Alert: Connect your LLM to external tools with Anthropic + MCP

  • Meta Is Delaying the Rollout of Its Flagship AI Model

TikTok AI Alive Brings Photos to Life by Turning Them into Videos

TikTok has launched AI Alive, its first AI-powered image-to-video feature, available exclusively through TikTok Stories. This tool allows users to transform static photos into animated, immersive short-form videos using AI-generated movement, effects, and sounds, making storytelling more dynamic and accessible even for those without editing experience.

Key Features:

  • Accessible via the Story Camera (blue "+" on Inbox or Profile pages).

  • Users select a photo and tap the AI Alive icon in the toolbar.

  • Videos are viewable on the For You, Following, and Profile pages.

  • Enhances photos with cinematic effects, such as moving skies, drifting clouds, or ambient sounds.

  • Aimed at boosting creativity and engagement for over 1 billion users.

AI: Your Career Catalyst, Not a Competitor

The narrative surrounding artificial intelligence (AI) often centers on job displacement. However, recent trends indicate a more nuanced reality. Lenny's latest jobs report reveals a surge in open positions for product managers (PMs) and engineers, reaching a peak not seen in the past 2.5 years. This uptick suggests that AI is not merely replacing jobs but is also generating new ones.

Companies are leveraging AI to unlock novel product opportunities. For instance, Notion has introduced an AI suite featuring meeting notes and enterprise-wide search capabilities, enhancing productivity and collaboration . Similarly, Duolingo has launched 148 new language courses more than it had developed in the previous 12 years combined by harnessing generative AI .

Despite these advancements, there's a skills gap. Transitioning from traditional Web 2.0 applications to AI-native products requires a new set of competencies. Recognizing this, we're launching a suite of courses titled "Building AI-Native Products" to equip PMs and engineers with the necessary knowledge. This program offers tactical, expert-led guidance on:

  • Developing AI-native products that address real user needs

  • Understanding evaluation metrics to enhance AI product performance

  • Making informed technical architectural decisions for generative AI initiatives

Google AlphaEvolve: The AI That Discovers Algorithms Humans Haven’t Thought Of

Google DeepMind just unveiled AlphaEvolve, a Gemini-powered coding agent that autonomously discovers algorithms more advanced than what we've crafted in over 50 years.

Yes, you read that right. AlphaEvolve rediscovered and even improved on legendary algorithms like Strassen's matrix multiplication something mathematicians have been optimizing since 1969.

But here’s where it gets even more exciting:

It doesn’t just stop at theory.
AlphaEvolve is solving real-world problems from data center scheduling to chip design and even improving large language model performance.

This isn’t just AI helping with coding.
This is AI redefining how we think about problem-solving itself.

🔍 Imagine pairing generative models like Gemini with automatic evaluation systems to evolve ideas at scale. No ego, no fatigue just raw exploration and improvement.

We’re witnessing the rise of AI systems that don’t just assist us they invent beyond us.
And this could be just the beginning.

📌 Curious about how it works? Dive into the official blog here

New Course Alert: Connect your LLM to external tools with Anthropic + MCP 🧠🔗

If you're building AI apps and tired of reinventing the wheel every time you want to connect your LLM to external tools, data sources, or prompts this course is for you.

I’m excited to share a brand-new course:
📚 MCP: Build Rich-Context AI Apps with Anthropic
Created in partnership with Anthropic and taught by Elie Schoppik, their Head of Technical Education.

Why take this course?

Most AI applications need context from web searches, codebases, PDFs, or APIs. But today, getting that context into your LLM often means writing a bunch of custom integrations.

Enter MCP (Model Context Protocol) an open standard developed by Anthropic that simplifies how LLMs interact with external tools, data, and prompts. It’s a game-changer for AI devs looking to build scalable, modular, and interoperable apps.

In this hands-on course, you’ll:

🔹 Build an MCP-compatible chatbot
🔹 Create & deploy your own MCP server with FastMCP
🔹 Connect to open-source tools like fetch (web content) or filesystem (file ops)
🔹 Plug your server into Claude Desktop and see it just work
🔹 Get a sneak peek at MCP’s future: multi-agent support, auth, registry APIs, and more

Meta Is Delaying the Rollout of Its Flagship AI Model

Meta has delayed the release of its flagship AI model, "Behemoth," due to internal challenges in improving its performance. Originally set for an April launch to coincide with Meta's AI developer conference, the release has been pushed to fall or later. Engineers are struggling to achieve significant advances over previous versions, raising internal doubts about its readiness and public release.

Despite Meta’s claims that Behemoth outperforms rivals like OpenAI and Google on some benchmarks, internal tests reveal performance issues and training difficulties. Executives are frustrated with the Llama 4 team’s progress and are considering management changes in the AI group.

This situation reflects broader struggles across the AI industry, with delays also affecting OpenAI's GPT-5 and Anthropic’s Claude 3.5 Opus. Meta has heavily invested in AI, planning up to $72 billion in capital expenditures this year, but staffing changes and credibility issues—such as submitting a benchmark-optimized model to a public leaderboard—are raising concerns about the company’s direction and transparency in AI development.