Welcome back! A reporter tested Amazon's Bee wearable (ICYMI: the AI wrist gadget Amazon acquired last year) and came away "both intrigued and slightly creeped out." Bee records, transcribes, and summarizes your conversations throughout the day.
It worked well for professional meetings, faithfully producing summaries broken down by topic. But the wearable needs expansive permissions: location, photos, contacts, calendar, notifications, even health data. And it’s all stored in the cloud. The reviewer's verdict: promising for work, but too invasive for personal life.
Would you wear an always-on recorder? Hit reply and share your thoughts.
Is Google About to Break the Web That Feeds It?
At I/O last week, Google unveiled the biggest redesign of search in over 25 years—and it's basically turning search into an AI engine by default.
A few of the big changes:
The search bar is becoming a full prompt box. Instead of typing a one-line query, you type out a full question, and you're essentially in AI Mode by default.
"Search agents" run in the background, continuously searching on a topic you care about and pinging you when something changes. Feels like Google Alerts, evolved.
Google can now write code right inside the search response. For example, ask "how do black holes affect spacetime?" and it builds an interactive visual with a slider you can drag around.
Here's why this is controversial: Say you search "best video games of 2026." Right now you get a small AI overview up top, but below it are organic links—real blog posts from real people who rely on you clicking through. Those sites are ad-supported, subscription-supported, or trying to grow a list or sell something. That click is how they survive.
But if Google scrapes those same blog posts, packs the answers into an AI overview, and nobody ever clicks through, that's a drastic cut in traffic to the very sites the answers came from. And they're rolling out the same idea to YouTube with an "Ask YouTube" feature that answers your question without you necessarily watching the video.
My POV: There's a real chicken-and-egg problem here. If people lose the incentive to create content because nobody visits their site anymore, they stop creating it—so where does Google pull its answers from then?
I honestly don't know how smart it is to remove that incentive. It also cracks open a door: a competitor (a Bing, or someone new) could swoop in and say "give us a prompt and we'll just show you organic links, no AI answer," and the people who miss the old Google might just migrate over.
— Matt
Pope Leo Warns AI Weapons Are ‘Beyond Human Control’
Via The Catholic Herald
Pope Leo released his first encyclical Monday—a nearly 43,000-word document focused on AI—warning that some autonomous weapons systems have advanced "practically beyond any human reach to govern them."
What he's calling for: The pope urged governments to slow AI development and implement robust regulation. He called for AI data ownership not to be left solely in private hands, for policymakers to protect workers' rights and keep children safe from the technology, and for AI companies to cool their competition.
"What is needed is a more active political involvement that is capable of slowing things down when everything is accelerating.”
On AI in warfare: Leo said any use of AI in warfare needs strict ethical guardrails and called it "not permissible" to entrust AI systems with lethal decisions. He warned that the ease of deploying autonomous weapons "makes war more feasible and less subject to human control."
Who was there: The document was released at a Vatican event where Anthropic co-founder Chris Olah also spoke, acknowledging that AI labs operate "inside a set of incentives and constraints that can sometimes conflict with doing the right thing." He thanked the pope for the outside scrutiny.
Why it matters: When the pope and an AI lab co-founder are standing together calling for regulation, it's a signal. The world's most influential voices—religious, political, technical—are sounding the alarm. The question is whether governance can keep pace with the technology.
Huawei Says It Can Match Cutting-Edge Chips Within 5 Years
Huawei announced Monday that it had achieved a breakthrough that would allow it to make chips equivalent to 1.4-nanometer processes by 2031, a direct challenge to US sanctions designed to cut China off from advanced semiconductor technology.
The new design: At a tech conference in Shanghai, Huawei's semiconductor president He Tingbo unveiled a new "LogicFolding" design that stacks traditional 2D circuits into 3D vertical structures. Instead of shrinking transistors, which requires lithography machines China can't access, Huawei is folding and stacking chips on top of each other.
The context: US sanctions since 2019 have largely blocked Huawei from global chipmakers. China's most advanced capability is currently around 7 nanometers, while Taiwan's TSMC—which makes chips for Nvidia—uses 2-nanometer technology and plans to hit 1.4 nanometers by 2028.
The bigger picture: The global AI arms race runs on chips. If Huawei’s announcement is any indication, the US sanctions designed to slow China down may have actually accelerated homegrown innovation. Whether Huawei can deliver on this timeline remains to be seen, but the ambition is clear.
AI-powered YouTube thumbnails
Via Thumbs.ai
Thumbs.ai generates YouTube thumbnails from text, titles, images, or video links. It analyzes top-performing channel styles to design thumbnails with high click-through rates.
How you can use it
Generate thumbnails instantly without design skills
Clone your channel's visual style for consistency
A/B test variations to find what performs best
Use the AI CTR rater to predict click-through rates
Pricing: Free and paid
Scientific figures from plain English
Via FigCanvas
FigCanvas turns prompts, methods text, or uploaded datasets into publication-ready illustrations, flowcharts, and data visualizations.
How you can use it
Build polished multi-panel figures for papers
Generate flowcharts and diagrams from text descriptions
Export in SVG, PNG, or PDF formats
Skip hours of manual formatting in design tools
Pricing: Free and paid
Jobs, announcements, and big ideas
Apple registers a new GenAI subdomain ahead of WWDC, signaling the long-rumored Siri overhaul in iOS 27.
Anthropic's Project Glasswing uncovers more than 10,000 critical software vulnerabilities in its first research update.
Microsoft AI launches MAI-Image-2.5, debuting at No. 3 on the Arena leaderboard among image generation models.
OpenAI is hiring a new chief marketing officer from ServiceNow.
Figure inks a deal with Catalyst Brands to deploy humanoid robots across the retail conglomerate's operations.
Universal Music Group and TikTok renew their licensing pact with new guardrails against unauthorized AI-generated music.
ElevenLabs ships Music v2 with sharper vocals and expanded multilingual support for AI-generated tracks.
ChatGPT can now see your bank account. Here's the quick breakdown:
That’s a wrap! See you Friday for more.
