The Devin hype was overblown. Techies knew the main challenge was the language model, likely simulated in their demo. They were essentially repackaging function calling capabilities of language models then as a novel product. Those discussing AGI often overlook that it's fundamentally about data patterns. Language models excel due to abundant text data, with code following for similar reasons. The core challenge remains data availability and quality
@deter34 ай бұрын
wish u can deliver Devin .
@yusefnathanson4 ай бұрын
The interviewer keeps saying Devin’s launch was a success. But it hasn’t launched yet. Why gaslight the audience? This trump-like hyperbole is bad for the whole AI industry.