Cal Newport has had enough. In a new essay published this week, the Georgetown professor and author calls out Silicon Valley for abandoning actual problem-solving in favor of hype cycles designed to enrich investors. He draws on Elizabeth Lopatto's recent piece in The Verge, which argues that tech companies used to build products people wanted. Now they build products and expect people to want them. The bandwagon list is familiar: NFTs, the metaverse, and now large language models. Lopatto's line is sharp: these technologies "are not built to really solve a market problem. They are built to make VCs and companies rich."Newport to AI: Figure Out What You Actually Do
Newport concedes that LLMs have more real utility than NFTs ever did in Newport to AI: Figure Out What You Actually Do, but he pushes back on the idea that this excuses the constant hype. Most people use ChatGPT as a verbose Google search, he writes. Maybe they format an itinerary now and then. Useful? Sure. Life-changing? Less so than the iPod was in 2003. The difference is that nobody was forced to endure daily panic about how the iPod would reshape civilization.
He also catches the media in a contradiction that matters for anyone tracking AI's actual impact. Over the past year, outlets confidently reported that AI was destroying entry-level hiring for college graduates. Then last week, new data showed the entry-level job market rebounding with projected hiring increases. The Wall Street Journal covered the positive numbers by suggesting AI was actually creating jobs. So which is it? The whiplash reveals how little anyone actually knows about how this technology affects employment.
The structural problem runs deep. VC firms operate on power law returns, where one massive hit pays for every other loss in the fund. This means investors can't afford to miss the next paradigm shift, even if normal people haven't asked for one. The result is a herd mentality that floods capital into whatever narrative promises 100x returns, consumer demand be damned. Newport's advice to AI companies is blunt: build something that noticeably improves people's lives, then talk. Until then, stop with the breathless announcements.