On‑Device AI Powers Gemini Nano, Core to Chrome Strategy
We’ve seen some questions about Gemini Nano on @googlechrome, so I want to clarify a few things. On-device AI is core to our developer & security strategy. 🧵
Stop Defaulting to Platforms; Choose Media Strategically
Media buying decisions should be strategically arrived at, but rarely ever are. Instead, channels typically get selected as a default, and then the strategy is fitted into them, and it's a problem across marketing. This quote from the nearly 30...
Kevin O'Leary Backs Utah AI Data Center Venture
#MrWonderful joins the #AI gold rush. #SharkTank's #KevinOLeary part of group pursuing 40000-acre prop in #UT to be anchored by a, you guessed it, #datacenter. #realestate #warehouse #MartySupreme #GreatSaltLake https://t.co/5TwcPOIBmB via @WSJ
Priority Network Keeps First Responders Connected During Congestion
Congestion doesn’t pause for emergencies. Having served as a public safety director, I know emergency responders depend on reliable communications. T-Priority dedicates priority access to first responder traffic so dispatch stays in control. @T_Priority Partner https://t.co/HB8vF8jR14
Google Cloud Transforms Into AI Hypercomputer for 2026
Is Google Cloud still "just another cloud"? Not anymore. The shift to the AI Hypercomputer marks a move from general infrastructure to a purpose-built, integrated AI stack. Essential reading for CIOs planning their 2026 roadmap. https://t.co/CqZ5slgQDX #CIO #AI #Cloud #DigitalTransformation #GoogleCloudNext
Even Smarter AI Can't Create Net Social Value
No matter how smart AI gets, the total social return to zero-sum activities is zero
Peptides Boost Hormones, Yet Trigger Serious Metabolic Side Effects
People mistakenly believe peptides are only good. Peptides can be bad, too. They can cause adverse effects. Some dangerous. I did a peptide experiment and measured its effects in my body. The results are complicated. I tried a peptide called CJC-1295....
Google Gemini's Lyria Generates Full Songs, Not Just Lyrics
Creating song lyrics with AI is so 2025. Now you can compose entire songs using the Lyria model in Google Gemini. It's pretty darn astonishing. Here's how it works, with sample tracks in a variety of styles: https://t.co/7pNLIXLd7b #gemini #lyria...
DARPA's Quiet Hybrid‑Electric Flying Wing Drone Takes Flight
DARPA’s XRQ-73 Hybrid-Electric Flying Wing Drone Has Flown The XRQ-73 was designed with a focus on very quiet, high-efficiency flight, and it has evolved since it was last seen in 2024. https://t.co/44mi9iXh5Y
Honest Feedback Loops Beat Fancy Eval Dashboards
"The teams that win the next phase of AI engineering won’t be the ones with the most elaborate eval dashboards; they’ll be the ones with the most honest feedback loops." https://t.co/yx5BRTSHCV < invest in evals, says @mjasay
Shorts Are My Top Growth Driver, Newsletter Coming Soon
Fascinating insights from my @YouTube channel: Shorts are WINNING for me, and I will be doing a deep dive on this in my newsletter very soon📈 https://t.co/81VS3AtmgC
Fear Grows When Anthropic Theory Might Be True
anthropic scares me but also they might be right and that scares me more https://t.co/Jyrg4FtNFh
Hamilton Unveils Integrated Glucose Sensor, Marks Decade Milestone
Hamilton Company has been coming to @SynBioBeta for years. This year, they showed up with something new. Tyler Schweder from Hamilton's process analytics team walked Gary George through their full upstream workflow: pH, dissolved oxygen, viable cell density, total cell...
AI Risk Stems From Human Recklessness, Not Innate Malice
First of 2 essays by @mboudry which I think nail the problem of AI existential risk: Neither malevolence nor lethal means to an end are natural products of intelligence, though they could arise if anyone was so stupid and reckless...
Musk Masters Scaling, Struggles with Frontier AI Breakthroughs
I don’t think I’ve seen this take before but I like it. Musk has been world-leading at compressing money, resources, and time to make “known/hard” things at scale—make an electric car, make batteries, make a cheaper bigger rocket, all...
Most AI Deployment Limits Lie Beyond the Chip
This is going to be an interesting report. Quick highlight: Our map identifies 25 constraints between announced AI capacity and billable compute. Ten can gate deployment outright, and most of those sit outside the chip itself.
Reliability Is the Core Challenge in Generative AI
Precisely what I have been saying since day 1: the challenge in generative AI is reliability. That was the challenge, and still is.
Guest Posts That Actually Expand Your Blogging Network
RT @RyanBiddulph Writing Guest Posts for Blogging that Actually Grow Your Network https://t.co/Fa7HOkHaR4 via @InspireToThrive
Germany's Tank Production Is Irrelevant; Drones Matter
Yes. Germany can make plenty of tanks but who cares? Tanks are obsolete. How many drones can Germany make?
1M Context Window Underdelivers, Slows After
what does this even mean? 1m context window hasn't been as great as advertised; claude clearly gets dumber and slower with 700k+ tokens. the only way to clean it up is to use /compact which defeats the purpose of the...
Texas Beats California in Clean Energy Thanks to Fewer Regulations
Everyone who cares about climate should understand this. Texas, with no pro-climate policies, has blown passed California in clean energy. In large part because Texas has less red tape and makes it easier to build.
Error Propagation Undermines Accuracy of Complex Autonomous AI
Errors in agentic networks multiply as they propagate. The more complex the multi-step process, the less accurate and more expensive fully autonomous AI becomes. That's the case for keeping humans in the loop. #AIAgents #RiskManagement https://t.co/iNfGKDSPeN
AI Compute Costs Equal XAI’s Quarterly Loss
This likely costs about ~$500m a month, ~$6bn a year to rent blended Which is about the run rate net loss of xai end of q1 Anthropics revenue run rate was $9bn end of 2025, $30bn a month ago
AI Search Success Requires Third‑Party Brand Visibility
B2B SaaS companies really nailing AI search strategy right now are not just thinking about owned content; they're thinking about how the brand shows up across third-party sites, too. IMO, it probably always should have been this way. https://t.co/oVIRa7IDLM
Invest in Reddit Now to Capture B2B Intent
The best time to invest in Reddit was last year. The second-best time is now. B2B buyers openly compare tools there every day. Install the Reddit Pixel. Retarget pricing-page visitors. Run always-on ads. Capture intent while competitors wait. https://t.co/OF7FNXFH4b
Claude Team Delivers Live Coding Session at 13:45 PT
Lots of good sessions with Claude team. including a Live coding session with Claude Code and Boris at 13:45 PT
Frontier AIs Bleed 33% in Two Weeks
The best models losing 33% of your portfolio value in 2 weeks: Take Alpha Arena, run by tech startup Nof1. It pitted eight major frontier AI systems — including Anthropic’s Claude, Google’s Gemini, OpenAI’s ChatGPT and Elon Musk’s Grok — against...
AI Lead-Gen Site Earns $1K Daily, No Dropshipping
Forget Dropshipping. This “Boring” AI Website Prints $1,000/Day (Feels Illegal) WATCH HERE 👉 https://t.co/2CH7VDOJv6 In this video, I break down how this AI website model works using a real example, a Dallas AC repair lead-generation site built with AI. I show...
Supreme Court Forces Apple to Reveal App Review Costs
The Supreme Court has considered Apple's delaying motion and found it unworthy. Now Apple will have to come clean with the District Court and show exactly what its costs are for reviewing apps with competing payment systems, so they can bill...
Customers Accelerate Orders to Dodge Price Hikes, Supply Risks
While the AI infrastructure business is certainly robust, something I've been suspecting and mentioned in all the manufacturing PMIs (broadly) is the pull forward of ordering to get ahead of price increases and/or supply chain issues. That goes too for...
Ilya's Concern Was Sam's Misconduct, Not AGI
100% stood the test of time: “What Ilya saw” was Sam’s bad behavior, not AGI.
AI Agents Now Retain Memory, Evolving Into Persistent Systems
Most AI agents forget everything after each task. That’s changing. Claude just introduced agents that learn over time. Dreaming builds memory from past sessions. Outcomes enforce quality with a grader loop. Multi-agent workflows handle complex tasks faster. Webhooks notify you when it’s done. This is a shift...
Musk‑OpenAI Case: Rare Factual Consensus, Ethical Question Remains.
Many trials feel like Rashomon with different witnesses. The amazing thing about Musk-OpenAI is how much agreement there has been (at least so far) on the facts. The question is really whether what OpenAI did is ok, not about whether...
LinkedIn Creator Partnerships Boost B2B SaaS Leads
B2B SaaS are getting smart about leveraging LinkedIn creator partnerships and seeing real impact on lead generation + sales. I did a deep dive on the data in today's newsletter: https://t.co/Gjlq6bTA8f https://t.co/Q5MXl083gP
Telesat's GEO Business Falls 25%, Loss Triples
Seems like a lot. "Telesat’s Geostationary Orbit (GEO) satellite business declined 25% year-over-year in the first quarter of 2026" Net loss for the quarter was CA$151 million ($111 million) compared to a CA$51 million ($37 million) loss in the prior year, https://t.co/cluTNPsoHJ
GPT‑5.5 Leads Child Safety Rankings, Anthropic Lags Behind
Child safety scores for GPT-5.5, Opus 4.7, and DeepSeek V4 are now live on KORA. A few interesting learnings: • GPT-5.5 achieved OpenAI’s highest child safety score to date: 75%, ranking 3rd out of 35 models • It's the first time Anthropic...
AI Compute Shortage Grants Providers Massive Pricing Power
One thing is for certain, the demand for AI compute resources far exceeds the supply, and probably will for the foreseeable future. Probably for a few years at least. Companies are going to have to compete for limited AI resources. They’re...
AMD Poised to Dominate Expanding Data Center CPU Market
Shared some thoughts on $AMD along with how we see the data center CPU TAM expansion playing out.
Early Manufacturing Rebound Fuels AI‑Driven Earnings Surge
We’re still in the early innings of a MFG PMI rebound, which is adding to the AI capex boom. We spent three years without a cycle, while AI spending was solely driving earnings. Now, we have more companies seeing +...
Anthropic Eyes Gigawatt Orbital AI Compute via Rumored Dario Centers
It is just a rumor that the SpaceX capacity Anthropic will use will be provided by satellites called: "Dario Centers." "As part of this agreement, Anthropic also expressed interest in partnering to develop multiple gigawatts of orbital AI compute capacity." https://t.co/qlH6StuyIB
Robinhood Offers Seniors Early Dividend Payouts via Gold
Robinhood's version of "get your paycheck early" for seniors. if you're a retiree and rely on dividends for your income, well, robinhood now gives them to you 1 month early... gated behind gold subscription.
AI Code Assistants Erode Developer Thinking and Planning
Still heavily using hermes + minimax M2.7 for daily code development largely in place of claude code w/ opus 4.7. I'm at 21M tokens so far, mostly because I wasn't using with work til I could get a license...
Never Share Financial Details Online—Avoid Scams
Please stop disclosing your financial information on social media. What is the purpose? Putting yourself and your money at risk of possible scammers is not a flex. Is a hazard.
MSPs and SMEs Get Pentagon-Level Threat Briefings
When the MSP and SME Security Vendor Reads the Same Threat Briefing as the Pentagon https://t.co/PNc7eAHlen
Open‑Source Community Launches Multica Autopilot Clone Locally
The open-source community is INSANE 🤯 ICYMI Claude Code routines dropped days ago, and this guy's already shipped a clone. it's called Multica Autopilot: Now you can run routines entirely locally using whatever agent you want (Opencode, Codex, Hermes, OpenClaw etc). repo in 🧵↓
Anthropic's Team Shows Ethical Rigor, Claude Likely Safe
“No one set off my evil detector. So long as they engage in critical self-examination, Claude will probably be good.”
Google Cloud IAM Gets Major New Features and Guardrails
All of a sudden, identity management is a vibrant and exciting space. We made a ton of @googlecloud IAM improvements lately across agent identities, gateways, guardrails, and more. Check out this recap: https://t.co/ShLXPI78Md https://t.co/C1IBJ1ylrN
Map Content, Build Pillars, Cluster, Link—Skip Random Blogs
Chill and Think Boomer Mike beats Huff N Puff Millenial Dwayne just for this one thing: No random blogs. Instead, Uncle Mike: - Map its content - Creates pillar pages - Builds supporting clusters - Connect everything with internal links No need to overcomplicate it.
AI May Outdiagnose Doctors, but Context Matters
A recent Science paper suggested doctors are now outstripped at diagnosis patients by AI - but there's more than meets the eye. My latest for @bmj_latest https://t.co/yE3r2q8U7U
Tree‑Climbing Pruning Robot Outpaces Monkeys
Faster Than a Monkey: This Tree-Climbing Pruning #Robot Is Wild by @IntEngineering #Robotics #Engineering #ArtificialIntelligence #Innovation #Technology https://t.co/O7ilwkze3q