In partnership with

Welcome back, AI enthusiast.

Google pushed AI deeper into search, and users immediately started downloading DuckDuckGo instead.

DuckDuckGo saw US iPhone installs jump 33% week over week, with visits to its no-AI search version surging too. Is Google's AI bet quietly costing it loyal users?

Today in the “Incredible AI” world:

  • DuckDuckGo Surges After Google AI Push

  • EAGLE 3.1 Solves the AI Attention Problem

  • Pope Leo XIV Meets Machine Creativity

  • Find Research-Backed Answers with AI

  • 5 Super-Useful AI Tools You Should Try

  • Top AI-Generated Image of the Day

Read Time: 4 minutes

- LATEST DEVELOPMENTS -

Google Pushed AI Hard, Users Pushed Back:

Google just made Search a whole lot smarter. Turns out, a lot of people weren't asking for that. DuckDuckGo is quietly picking up the tab for Google's ambitious AI overhaul.

Things to Know:

  • The Numbers That Matter: iPhone installs of DuckDuckGo in the US jumped 33% week over week after Google I/O. That's nearly double the 18.1% global growth rate the company saw overall.

  • The AI-Free Option Wins: Visits to noai.duckduckgo.com also surged at a 22.7% average weekly growth rate. People aren't just downloading the app. They're specifically choosing the no-AI lane.

  • Why the US Specifically: DuckDuckGo noted that US growth ran at multiples of the international rate. Google's AI Search overhaul was announced as a US-first rollout, which likely explains why the backlash is concentrated there.

  • The Irony Worth Noting: DuckDuckGo itself has AI features. The company just keeps them optional, which is exactly the approach that's winning users over from Google right now.

Google called its AI Search update "the biggest upgrade in over 25 years." Some users called it the last straw. The backlash was real enough to send DuckDuckGo installs spiking across the US.

New Algorithm Boosts LLM Performance:

Your AI model's tiny helper had a focus problem. It kept getting distracted by its own outputs instead of the original context. EAGLE 3.1 just fixed that, and the speed gains are worth paying attention to.

Things to Know:

  • The Problem It Solves: LLMs use a small draft model to predict tokens before the big model checks them. As guesses go deeper, the drafter loses focus and slows everything down.

  • Two Fixes, Real Results: EAGLE 3.1 adds FC normalization and post-norm hidden-state feedback. These two changes keep the drafter stable and accurate across longer speculation depths.

  • The Speed Numbers: On benchmarks with Kimi K2.6, EAGLE 3.1 delivers 2.03 times higher per-user output throughput at low concurrency. In long-context tasks, it achieves up to 2 times longer acceptance length versus EAGLE 3.

Speculative decoding has always been a clever trick for squeezing speed out of large models. EAGLE 3.1 fixes the part where that trick quietly broke down in production. If you run LLMs at scale, this one is worth testing.

Bloomberg: "No reliable safe havens." Billionaires have been investing elsewhere. Here's how to get in.

Bloomberg's Marcus Ashworth wrote plainly recently: "No more reliable safe havens."

After all, the S&P fell over 7% from the February peak. Bonds, even with less risk, are barely keeping pace with inflation.

So-called "diversified" portfolios have gotten hit from multiple directions.

Meanwhile, the world’s wealthiest have been setting records in another asset class.

Circumstances are always unique, but after the dot-com bust, it grew roughly 24% annually for a decade. After 2008, roughly 11% annually for 12 years.

Blue-chip art.

Why? It trades globally in multiple currencies, has scarce supply, and has shown near-zero correlation to equities since 1995.*

With Masterworks, 70,000+ investors allocated $1.3B fractionally across 500+ artworks featuring Banksy, Basquiat, and Picasso.

Accredited? You can invest in a diversified portfolio of postwar and contemporary art alongside two other real assets. From 2017-2025, the mix would’ve beat the S&P 500 by 3.1x.

See if you can improve your portfolio performance all in one diversified strategy.

*According to Masterworks data. Investing involves risk. Past performance is not indicative of future returns. Important Reg A disclosures: masterworks.com/cd

The Pope's AI Warning May Be AI-Written:

Pope Leo XIV just published a 42,000-word encyclical warning the world about AI. Then AI detectors flagged chunks of it as AI-written. You genuinely cannot make this stuff up.

Things to Know:

  • What the Encyclical Says: Magnifica Humanitas warns governments and companies that AI must serve humanity, not replace it. It took a strong stand against letting machines define what it means to be human.

  • What the Detectors Found: When The Verge ran 2,000 words through Pangram, the tool flagged 46% as AI-generated. One chapter scored as high as 62% AI-written in a separate analysis.

  • The Paragraph That Stings: One flagged section read that no computer "can create a heart that gives itself." The irony of that specific line being AI-written is the kind of thing that practically headlines itself.

  • What This Actually Means: Pangram flags patterns, not proof. The Vatican likely used AI as a drafting or translation aid. That is still a long way from fully outsourcing a moral document to a chatbot.

The Vatican hasn't commented on whether AI assisted the writing. Encyclicals are typically team efforts, not solo papal manuscripts. Still, warning the world about AI while quietly using it to do it is a story that will follow this document for a very long time.

- TRY THIS -

How to Search 200M Research Papers Using AI:

When you want to back up a claim with actual research, the usual process is painful. You open Google Scholar, dig through abstracts, hit a paywall, and spend an hour finding one paper that half-answers your question. Most people just skip the research entirely and wing it.

Consensus is an AI-powered search engine built specifically for scientific literature. It searches over 220 million peer-reviewed papers and uses AI to summarize, synthesize, and surface what the research actually says, with citations attached to every claim.

Here's how to use it:

  • Sign up and run your first search: Create a free account and type your question in plain language. Something like "does intermittent fasting improve cognitive performance" works just as well as a formal academic query. Consensus searches the full database and returns the most relevant papers ranked by relevance, recency, and citation count.

  • Read the AI summary at the top: Instead of opening ten papers yourself, Consensus generates a synthesized summary of what the top results collectively say. Every claim in the summary is tied directly to a source paper, so you can verify anything instantly.

  • Check the Consensus Meter: This feature shows you at a glance whether the research broadly agrees, disagrees, or is mixed on your question. It is particularly useful when you want to know how settled a topic actually is before you write or talk about it.

  • Use Deep Search for complex questions: For multi-layered research questions, Deep Search runs up to 20 targeted searches across the database, reviews over 1,000 papers, and produces a structured report similar to a literature review, complete with an introduction, methods, results, and discussion sections.

  • Save papers to your library: Papers you want to revisit can be saved into custom collections inside My Library. For open access papers, you can download the full PDF directly from Consensus without leaving the platform.

If you create content around AI, health, productivity, or any topic where citing real research adds credibility, this is a significantly faster way to find it than traditional academic search tools.

Trade from the subway, golf course, or toilet

Liquid enables users to go long or short on any market from your phone or desktop 24/7.

No matter what time of day it is or where you are, you can monitor the situation with Liquid.

- BEST AI TOOLS -

Visily: Turn screenshots and sketches into editable UI designs using AI assistance.

Kili Technology: Improve AI training data with advanced labeling and model feedback tools.

QuickVid: Automatically generate short-form videos from scripts and prompts.

Mentionlytics AI: Track brand mentions and analyze sentiment using AI-powered monitoring.

NeuralFrames: Create AI-generated music videos synced to your audio tracks.

- DAILY POLL AND RESULT -

Today’s Poll:

Q) Should AI help plan marketing campaigns?

Login or Subscribe to participate

Vote and find out the result tomorrow.

Yesterday’s Result:

Q) Is AI helping introverted founders pitch better?

  • A) Yes, confidence support - 53% 👑

  • B) No, authenticity matters - 47%

The Best New App for Reading Newsletters:

Reading newsletters in the inbox is frustrating - it is noisy and easy to lose control of subscriptions.

That’s where Meco comes in as a savior.

Meco is a distraction-free app for reading newsletters outside the inbox. The app is packed with features designed to supercharge your learning from your favorite writers.

With Meco, you can enjoy your newsletters in an app built for reading, while giving your inbox space to breathe.

Add your newsletters in seconds and liberate your inbox today!

- AI IMAGE OF THE DAY -

Happy Chipmunk Waters Plants With Care:

Prompt:

"A cute chipmunk wearing a tiny straw hat waters miniature tomato plants on a sunny windowsill using a thimble as a watering can. Bright sunlight, floating water droplets, and a joyful wholesome atmosphere create a warm realistic scene. Highly detailed fur, vibrant green leaves, tiny tomatoes, and soft natural lighting add charm and realism."

Try this prompt in any decent AI image generator and let me know your result.

- DO YOURSELF A FAVOUR -

If this email landed in your Promotions or Spam tab, move it to your Primary inbox. Takes 2 seconds, and you will never miss an edition again.

I put this together every day so you do not have to spend your morning hunting down AI news yourself. Would be a shame if it ended up buried under discount codes and bank alerts.

Keep Reading