4 ms·

Anthropic’s $5B, 4-year plan to take on OpenAI

• vagabund 4y ago
  > In the deck, Anthropic says that it plans to build a “frontier model” — tentatively called “Claude-Next” — 10 times more capable than today’s most powerful AI, but that this will require a billion dollars in spending over the next 18 months. > Anthropic describes the frontier model as a “next-gen algorithm for AI self-teaching,” making reference to an AI training technique it developed called “constitutional AI.” Anecdotally, Claude is pretty amazing. I default to it over GPT-4 for everything other than coding questions. > “These models could begin to automate large portions of the economy,” the pitch deck reads. “We believe that companies that train the best 2025/26 models will be too far ahead for anyone to catch up in subsequent cycles.” A once unified set of researchers operating under a nonprofit committed to safety, then fissuring over philosophical differences, and now racing each other to commercialize AGI is uncomfortably close to a sci-fi plot. All in all I'm excited though. There's a lot in this article.