⚠️ AI Power, Peril, and the Path Forward.

Plus: Artificial Analysis State of AI.

In partnership with

Get in Front of 50k Tech Leaders: Grow With Us

Midweek, and the AI headlines keep rolling. Smart glasses may soon be on every face—essentially putting your smartphone in front of your eyes 24/7, while collecting data non-stop for tech giants. Google just dropped a wave of new AI features, but with innovation comes risk. A biologist weighs in on the dangers, and we break down the latest State of AI from Artificial Analysis. Let’s dive in.

  • AI Power, Peril, and the Path Forward.

  • 🧰 AI Tools

  • Google I/O 2025 Highlights

  • 🧠 Learning Corner - Helsinki University

  • Artificial Analysis: State of AI.

  • 📰 News and Trends.

  • Google Enters the Smart Glasses Game.

StartEngine’s Blockbuster Year: How You Can Join the Action Before Our June Deadline

StartEngine is the platform allowing accredited investors to gain exposure to some of the world’s most coveted private companies like OpenAI, Perplexity, and Databricks — without paying millions.

Not surprisingly, they just posted new blockbuster financials:

📈 Monster (and record) Q4 revenues
📈 Revenue doubled year-over-year ($23M -> $48M)

They’ve been building something big, and the results speak for themselves.

The even better part? The window is open (but closing soon) for you to join their latest funding round. Over 50,000 have invested $84+ million in StartEngine — and now you can get in on the action before this round closes next month. Investments start as low as $500.

Reg A+ via StartEngine Crowdfunding, Inc. No BD/intermediary involved. Investment is speculative, illiquid & high risk. See OC and Risks on page.

AI Power, Peril, and the Path Forward.

In a recent talk, evolutionary biologist Bret Weinstein compared AI’s rise to the invention of writing, praising its near-infinite potential but warning of even greater risks.

He highlighted AI agents' growing autonomy and ability to solve complex problems, but cautioned that without ethical safeguards, miscommunication or abuse could lead to catastrophic outcomes.

Weinstein urges humanity to proceed with extreme caution, calling for robust frameworks to guide this “biblical-scale” transformation of human capability.

🌐 Other Tech news

  • Can Substack be the intellectual engine of the 21st century?

  • How 3D printing is personalizing health care.

  • Waymo says it reached 10 million robotaxi trips, doubling in five months

  • The democratic party is desperate for a digital strategist to help the party capture the digital mojo that helped President Trump win.

  • Upway aims to capitalize on the surging popularity of e-bikes in the US

  • Nvidia CEO says US export controls on AI chips are a 'failure'

Google I/O 2025 Highlights

Google just went all in on AI — again. Here's what’s new:

  • 🤖 Gemini AI replaces Assistant with live vision + voice + task handling

  • ✉️ Smart Gmail replies in your writing style

  • 🎨 Imagen 4, Flow & Jules bring high-quality AI image, video, and code generation

  • 🔍 Search gets AI Mode with chatbot answers + virtual try-ons

  • 🕶️ Android XR glasses go live with AR overlays (texts, maps, translation)

  • 🌍 Fire-detecting satellites + drone delivery for disasters

🧰 AI Tools

Download our list of 1000+ Tools for free.

Google Enters the Smart Glasses Game.

After years of Meta dominating the smart eyewear scene with its Ray-Bans, Google has just unveiled sleek new AI-powered glasses, developed in collaboration with Samsung, Warby Parker, and Gentle Monster.

Unlike the clunky Google Glass of the past, this stylish prototype features Gemini-powered AR, displaying maps, texts, emails, and photos right on the lens.

Meta’s Ray-Bans can’t do that yet, so Google may leap ahead in the race for AI glasses. No release date or price has been announced yet, but 2025 is shaping up to be the year for smart eyewear.

Artificial Analysis State of AI.

Q1 2025 Highlights.

1. AI Intelligence is Evolving Fast

  • Reasoning models—LLMs that “think” before answering—now dominate the intelligence frontier.

  • OpenAI’s o4-mini (high) and Google’s Gemini 2.5 Pro lead in accuracy and reasoning depth.

  • Open-weight models like DeepSeek R1 and Llama 3.1 are closing the gap on proprietary models.

2. Reasoning Comes at a Cost

  • Reasoning models generate up to 10x more tokens and take 5x longer to respond vs. non-reasoning peers.

  • Developers must now evaluate token usage per task, not just price per token.

3. Inference is Cheaper & Faster

  • Inference for GPT-4-level intelligence is now 100x cheaper than when GPT-4 launched in 2023.

  • This is driven by smaller MoE models, speculative decoding, and new hardware (e.g. Groq, Cerebras, SambaNova).

4. China Is Rising

  • US labs still dominate reasoning models, but China leads in non-reasoning performance, with models like DeepSeek V3.

  • New challengers like Seedream and HiDream are shaking up image and video generation.

5. AI Agents Are Getting Real

  • Agentic systems now automate coding, research, UI control, and customer support.

  • Early 2025 apps are chaining 20x more model calls per task than last year.

6. Multimodal AI Surges

  • OpenAI’s GPT-4o now leads in image quality (ELO score: 1152).

  • Video generation competition heats up with Google’s Veo 2 overtaking OpenAI’s Sora.

  • ElevenLabs’ Scribe leads in speech-to-text accuracy.

Subscribe to keep reading

This content is free, but you must be subscribed to Yaro on AI and Tech Trends to continue reading.

Already a subscriber?Sign in.Not now

Reply

or to participate.