A Zoom Bug Let Anyone Take Over Your Device. It Took One Researcher 20 Prompts to Find.

AI is changing how people work, build, and make money.

If you're only using AI for simple prompts, you may be missing hundreds of opportunities to turn today's tools into something more valuable.

Explore 200 curated AI income ideas for beginners and pros, including opportunities you can start this week and practical applications for today's top AI tools.

TODAY IN AI

1  Alibaba's Qwen claims 3 billion downloads over the past six months, well ahead of Alphabet and Meta in the open-weight race.

Alibaba's Qwen family of models has surpassed 3 billion downloads, by the company's own count, putting it well ahead of the two companies leading America's open-weight push. Open-weight models are increasingly closing the performance gap: Alibaba's latest, Qwen3.8-27B, reportedly topped Anthropic's Opus 4.6 on several benchmarks, per Alibaba's own internal evaluations. Read more.

2  Anthropic's CEO publicly rejected the idea that AI access has to be either fully concentrated or fully distributed.

In two rare social posts, Dario Amodei argued for a middle ground between those two extremes. He also attributed the negative public sentiment around AI to a broader distrust in institutions, rather than to his own public statements, which he characterizes as balanced. For sentiment to shift, he says, AI needs to actually deliver on its promised benefits rather than argue its case rhetorically.

3  OpenAI's executive exodus continues as it prepares for an IPO, with its Chief Revenue Officer now heading out the door.

OpenAI's CRO Denise Dresser announced she is leaving the company, joining a wave of high-profile departures that includes Fidji Simo and Brad Lightcap. CFO Sarah Friar and President Greg Brockman reportedly met with investors Friday to discuss the lab's upcoming IPO, filed confidentially in June, which still has no set timeline. A senior sales executive walking out the door in the middle of investor meetings is not the optics a company usually wants heading into a public listing. Read more.

FROM THE FRONTIER

The vishing problem. Voice phishing attacks are rising, and today's voice models make them dramatically easier to pull off. Current models mimic human speech to a startling degree, complete with natural pauses and stutters. Bland AI and GPT-Live are both genuinely useful tools with real applications, but the same realism that makes them useful also hands scammers a powerful new way to coax information out of victims. Top financial firms including Citadel, Two Sigma, and Point72 have already been targeted, and they will not be the last.

The bigger shift. Voice cloning is one specific attack vector. The broader trend is that AI has made it dramatically easier to find and exploit vulnerabilities across enterprise software people use every day. Researchers recently uncovered a critical flaw in Zoom, a platform used by 70% of the Fortune 100, which puts the scale of exposure in real terms.

Zoomsday. The bug let anyone in a Zoom meeting take control of another participant's device without their knowledge. Finding and exploiting a flaw like this used to require deep specialized expertise, often weeks of dedicated work. This time, it took less than 24 hours and fewer than 20 prompts using publicly available AI models. Read the full story behind Zoomsday.

What this actually means. The technical skill required to find a critical vulnerability in widely deployed enterprise software has collapsed from a specialized expert skill set to twenty prompts and a day. That collapse cuts both ways: it makes it faster to find and patch flaws, and it makes it faster for anyone with bad intentions to find them first. The security industry is not equipped, in either headcount or process, for a world where the barrier to finding a Zoomsday-class bug drops this low this quickly.

PRESENTED BY MINDSTREAM

Turns out, AI is good for more than writing emails and generating LinkedIn posts. This guide is packed with 200+ actionable ways to build income with AI.

Inside you'll discover:

AI is changing how people work and how people earn, far beyond the simple “write me an email” prompt. Discover the possibilities and download the free guide to start cashing in today.

Get The Free Guide

IN THE KNOW

🤔 Too much? One person invited an unlikely companion to watch the solar eclipse with them: Claude, complete with a small body. The sentimental post sparked a debate, with many saying it is a clear example of AI psychosis (3.5M views).

🤯 Cursor acquisition. Cursor is nearly unrecognizable since its initial launch. A four-year-old Hacker News post has resurfaced showing Cursor's humble beginnings, now that the startup has officially been acquired by SpaceX for $60B (8.5M views).

🤣 ChatTJB. San Francisco residents thought they were using a new AI tool. Turns out their messages were just going to some guy named Tucker who put up a fake billboard. Read the wild story (4.5M views).

🤖 Agent meetings. This Slack thread is full of agents, but feels surprisingly human. Why? Because one agent made up an excuse for why it was running behind on work, entirely unprompted (11,000 upvotes).

🔏 Watermarking. Anthropic's decision to watermark Claude content has become a genuinely hot-button issue. The lab posted a blog article further explaining how it works, while one engineer created a visualization that went viral (1M views).

PROMPT STATION

For [INSERT DETAILS] under existing customers, include real patterns you have noticed, even informal ones: "our best customers tend to come from referrals and close within two weeks" gives the model something concrete to work from. For [INSERT MARKET], go narrower than feels natural: "ops leads at Series A-C logistics startups" produces sharper output than "B2B software buyers." The red flags section is worth reading closely before the rest, since it tells you who to disqualify early rather than who to chase, which is often the more expensive mistake in sales.