AI-assisted cheating proves ineffective for students

The controversial marriage of cheating and AI reveals itself to be more marketing than substance as Columbia University student Roy Lee pivots from scandal to startup. After being suspended for creating AI software to automate coding interviews, Lee has secured $5.3 million in seed funding for Cluely, promising real-time AI assistance during virtual interactions. Early reviews suggest the product falls significantly short of its ambitious claims, highlighting the growing gap between AI hype and practical application in emerging consumer products.

The big picture: Recent college student Roy Lee transformed academic controversy into venture capital after being suspended from Columbia University for creating an AI app that completed coding interviews on his behalf.

• Lee raised $5.3 million in seed funding to launch Cluely, positioning the tool as helpful tech rather than cheating software, similar to how calculators and spellcheck are now normalized.
• Despite viral attention and bold promises, early reviews from technology journalists suggest Cluely performs poorly in actual testing scenarios.

How it works: Cluely claims to provide real-time assistance during virtual meetings by reading screens and interpreting audio without being detected by monitoring software.

• The tool promises to deliver background information about people in calls as if users had done extensive research, essentially eliminating the need to memorize information.
• Cluely offers a limited free version that Lee claims has 80,000 downloads, while the full version costs $20 per month and reportedly has 700 paid subscribers.

Marketing versus reality: Lee’s promotional efforts include an advertisement showing him using Cluely through smart glasses to misrepresent himself on a date, despite the product being a desktop application.

• Reviewers from The Verge experienced significant technical issues, with the application failing to understand context and responding too slowly for real-time conversations.
• 404 Media’s testing revealed similarly disappointing results, describing Cluely’s output as generic ChatGPT responses that weren’t delivered quickly enough to be useful.

Why this matters: Cluely joins other recent AI hardware products like Humane’s AI Pin and the Rabbit r1 that have fallen short of their ambitious marketing claims, highlighting the persistent gap between AI hype and practical consumer applications.

• While AI excels at tasks like document summarization and note-taking, real-time conversational assistance remains technically challenging despite growing investment.
• The product’s limitations suggest that traditional preparation methods—like spending a few minutes researching someone before a meeting—remain more effective than current AI alternatives.