AI agents introduce a new security problem: software that can take action, not just generate answers.
The critical question is no longer only whether an output is correct. It's whether the system has the right permissions, safeguards and audit trail before acting.
Alex retweeted
I'm so sorry Google. I didn't recognize your game.
To be fair, you did not indicate at any point that you still had game like this. I'll take the L either way. I'm just pumped that we finally have a good Gemini model.
The most underrated AI metric might be cost per successful task.
A cheaper model that needs five retries isn't necessarily cheaper. A powerful model that requires constant human correction isn't necessarily productive.
Benchmarks matter. Workflow economics matter more.
1
11
AI is compressing the time attackers need to find and exploit vulnerabilities. That's a structural change in cybersecurity economics.
Defenders can't rely on adding more analysts to the same manual workflows. Detection, patching and response need automation too.
18
An AI demo can look flawless and still fail in production. The real moat is increasingly in data quality, evaluation systems and workflow integration. Model performance gets attention. Operational reliability earns the budget.
2
2
75
Alex retweeted
How many Michael Jackson level talents do you think lived in the middle ages and were just huge legends to their village of 400 people in Spain
AI adoption is entering a harder phase. Buying model access is easy. Redesigning workflows, measuring reliability and proving ROI are not. The companies that solve those operational problems will capture more value than those chasing every new benchmark
1
48
🏎️ AI FINANCIAL TELEMETRY
Gartner reviewed nearly 1,000 case studies and found fewer than 10% of agentic AI deployments delivering robust ROI.
The technology is moving fast. The business case is still catching up.
Are companies scaling too early?
1
31
Bitcoin is becoming a bipartisan conversation. The real test? Whether governments can handle its volatility as well as they embrace its potential. 👀
Florida elected a Republican governor who backed a U.S. Bitcoin reserve. Texas elected a Democratic senator who voted to create one for his state. The case for holding Bitcoin crosses party lines.
1
57
🏎️ TEAM RADIO: AI EDITION
Engineer: “Our new model beat every benchmark.”
Team principal: “And the customers?”
Engineer: “We're checking.”
Team principal: “Box. Box. Box.”
AI racing has plenty of fast laps. Someone still needs to sell the tickets.
22
🎃 AI HORROR INDEX
👻 Ghost: Investors disappearing after your pitch
🧛 Vampire: Cloud bills draining your runway
🧟 Zombie: AI startups with no paying customers
💀 Final boss: The CFO asking when you'll be profitable
What's the scariest thing in tech?
16
⚽ THE BENCH PLAYER EFFECT
Every championship team has someone who barely makes the highlights but changes the game when called on.
Tech is the same. We celebrate founders and CEOs, but the quiet engineers fixing impossible problems deserve their own trophies.
Whos underrated?
10
Alex retweeted
one of the biggest questions around AI isn't whether companies want more computing power.
it's how much capital the industry needs before the spending turns into durable returns.
data centres require chips, electricity, cooling, land and financing. those commitments don't disappear if demand grows more slowly than expected or the cost of borrowing rises.
that creates a useful distinction for investors: companies selling essential infrastructure today may have a different risk profile from companies betting on profits several years into the future.
both can benefit from AI, but the timing and certainty of their returns are different.
if you had to choose one part of the AI supply chain to invest in, would you pick chips, cloud infrastructure or the software customers actually use?
Tech companies slapping "Genie 4" on another video generator to hype up AI world models while users just want software that doesn't hallucinate basics is pure VC bait.
This is Genie 4, our most realistic and interactive world model yet. 🌐
Explore dynamic environments in real time to see how the scene responds, sounds, and how the people in it move, react, and speak.
27
Legacy financial institutions blaming victimized customers for "phishing" the second their own outdated security infrastructure gets compromised is classic corporate gaslighting.
Someone hacked my uncles Charles Schwab account and sold 49k of his stocks on Friday. Luckily he caught it on Monday morning and they were able to freeze the funds in time. Charles Schwab is trying to tell him that he must have shared his password or fell for a fishing
scam. Then the thieves tried to transfer the remaining funds to an account in Palm Springs. Looks like Schwab has been breached. And they are blaming their customers.
14
⚽ THE BENCH PLAYER EFFECT
Every championship team has someone who barely makes the highlights but changes the game when called on.
Tech is the same. We celebrate founders and CEOs, but the quiet engineers fixing impossible problems deserve their own trophies.
Whos underrated?
9
⚽ THE AI TRANSFER WINDOW
AI companies hiring top researchers feels like football deadline day.
Record contracts. Surprise departures. Rival teams stealing talent.
Only difference? Nobody's holding up a scarf for a machine learning engineer.
Yet. 👀
4
4
68


















