The top 3 tech drops and AI breakthroughs you need to know

🤖 Farmer loses entire crop after trusting bad AI advice
A Chinese farmer trusted an AI chatbot's herbicide advice, and it wiped out his entire sesame crop overnight. Wu, 67, sprayed his 24.7-acre field in Chuzhou with a chemical mix the AI recommended, one that happened to kill sesame along with the weeds, since sesame is itself a broadleaf plant.

🤖 Google's Gemini hits 1 billion monthly users
Gemini has surpassed 1 billion monthly active users, making it Google's 14th product to hit that mark and putting it on par with ChatGPT. Google says 63% of Gemini users interact via voice, the app generates +150 million images daily, and it now has +100 million active users on iOS alone.

🤖 Executives are cloning themselves with AI, risks and all
Execs including Uber's Khosrowshahi and Meta's Zuckerberg have begun deploying AI clones of themselves to handle presentations and routine business tasks. But the limits already show: An influencer's clone earned $72,000 in its first week before veering into unsolicited improper conversations.

Trends, fresh research, and sharp tips to fuel your career

📌 Anthropic's new Claude watermarks are sparking user backlash
Anthropic has begun embedding invisible watermarks in Claude's text output to comply with the EU AI Act's rules, sparking backlash from users who say it exposes work they'd copy-pasted into articles. Others pushed back, arguing the watermarks just make AI content detectable, not claim credit for it.

📌 Federal workers can use TikTok on work phones again
Federal employees can once again install TikTok on their government-issued phones after the Trump administration lifted the 2022 ban, according to an Office of Management and Budget memo. Several federal agencies have already launched official TikTok accounts.

📌 Everyday investors risk losing their information edge
Retail investors just became a bigger force in the stock market than ever, and now 3 changes threaten to leave them in the dark: the SEC letting report earnings 2x/year, a tight-lipped new Fed chair, and Trump Media's new $60,000-a-month API selling faster access to market-moving posts.

Accounts Receivable Analyst | Full Time | Costa Mesa, CA | In Person
Senior Network Engineer | Full Time | Columbus, IN | In Person
Endpoint Administrator | Full Time | Greene, NY | In Person
UX Architect | Contract | IBOTIX | Remote

AI is remarkably good at sounding certain. It delivers answers in complete sentences, with confident tone, at impressive speed. That combination is deeply persuasive, which is precisely what makes it dangerous when the answer is wrong and nobody checks.

There is a name for what happens when people stop checking. Automation bias is the tendency to blindly approve machine outputs without independent verification, and instead of acting as vigilant safeguards, human operators risk becoming rubber stamps. The pattern is already well documented across medicine, law, and education. In each field, the problem is that the person using AI stopped thinking critically the moment it offered a response.

The real-world consequences are not hypothetical. In 2025, a US lawyer was sanctioned by a court after submitting a legal brief containing fake AI-generated case citations, with a law clerk having used ChatGPT to draft the petition without any independent verification. A doctor in a separate study approved AI-drafted patient messages without reading them properly. These are failures of habit, specifically the habit of verifying before trusting.

What makes this pattern so persistent is something researchers call cognitive surrender. Researchers define cognitive surrender as adopting AI outputs with minimal scrutiny, thereby overriding both intuition and deliberation, with a Microsoft Research study finding that confidence in AI was among the strongest predictors of whether knowledge workers engaged in critical thinking at all. The more people trust the tool, the less they interrogate its output.

The long-term cost goes beyond individual errors. Stronger AI dependency is associated with cognitive offloading and weaker critical thinking, with students who frequently rely on AI for explanations and solutions more prone to accept recommendations with limited verification and reduced analytical initiative. Skills atrophy when they stop being used. Judgment weakens when it is routinely outsourced. The cumulative effect of countless small decisions to trust AI without questioning it is a slow erosion of the very cognitive capacity needed to catch its mistakes.

None of this argues against using AI. It argues against using it passively. Automation bias is a persisting risk across high-stakes settings. Potential mitigations include helping users form accurate expectations of how a system performs and addressing the cognitive shortcuts that make unquestioned trust so appealing in the first place.

The tool is only as reliable as the judgment brought to it. Keeping that judgment sharp, rather than handing it over, is the whole point.

Last week’s winning choice:
Would you dine at a 100% AI-run restaurant?
🍽️ JUST FOR FUN: I'd try it as a gimmick, but I prefer real chefs (36.25%)

Got a friend who values AI, tech, and everything in between? Forward them today's read!