Anthropic’s $5 Billion 4-Year Plan to Address OpenAI

0
163

Start up AI research Anthropic aims to raise as much as $5 billion over the next two years to beat rival OpenAI and enter more than a dozen major industries, according to corporate documents obtained by gotechbusiness.com.

A pitch deck for Anthropic’s Series C fundraising round reveals these and other long-term goals for the company, which was founded in 2020 by former OpenAI researchers.

In the deck, Anthropic says it plans to build a “frontier model” — tentatively dubbed “Claude-Next” — 10 times more capable than today’s most powerful AI, but will require $1 billion over the next 18 months to do so. will be .

When contacted for comment, an Anthropic spokesperson said, “We are planning additional product announcements and will talk about them shortly.”

The information reported in early March that Anthropic attempted to raise $300 million at a valuation of $4.1 billion, bringing the total to $1.3 billion. The deck confirms that target number, though only half was raised at the time the document was created by a “confidential investor.”

Anthropic describes the frontier model as a “next-gen algorithm for AI self-learning,” referring to an AI training technique it developed called “constitutional AI.” At a high level, constitutional AI is trying to provide a way to do that Aligning AI with human intentions — make systems respond to queries and perform tasks using a simple set of guiding principles.

Anthropic estimates that its boundary model will require FLOPs or floating point operations on the order of 10^25 – several orders of magnitude larger than even today’s largest models. Of course hhow this translates to computation time depends on the speed and scale of the system performing the computation; Anthropic implies (in the game) that it relies on clusters with “tens of thousands of GPUs”.

This groundbreaking model could be used to build virtual assistants that can answer emails, conduct research, and generate art, books, and more, some of which we’ve already had the opportunity to sample with GPT-4 and other major language models.

“These models could start automating large parts of the economy,” the pitch deck reads. “We believe that companies training the best 2025/26 models will be too far ahead to catch up in the next cycles.”

The frontier model is the successor to Claude, Anthropic’s chatbot that can be instructed to perform a range of tasks, including searching documents, summarizing, writing and coding, and answering questions on certain topics. In these ways it is similar to OpenAI’s ChatGPT. But Anthropic claims that – thanks to constitutional AI – Claude is “much less likely to produce harmful output”, is “easier to talk to” and “more controllable”.

Anthropic released Claude commercially in March after a closed beta late last year, initially allowing about 15 partners access. It counts among its beta users and potential customers the following industries (with the asterisk indicating that a human is in the loop to oversee the model):

  • Summary and analysis of legal documents*
  • Medical patient records and analysis*
  • Customer service emails and chat
  • Coding models for consumers and B2B
  • Productivity related search, document editing and content generation*
  • Chatbot for public questions and answers and advice
  • Search with natural language responses
  • HR tasks such as job descriptions and interview analysis*
  • Therapy and coaching
  • Virtual Assistants*
  • Education at all levels*

Dario Amodei, the former vice president of research at OpenAI, launched Anthropic in 2021 as a public-benefit company, bringing with him a number of OpenAI employees, including OpenAI’s former policy chief Jack Clark. Amodei split from OpenAI after a disagreement over the company’s direction, namely the startup’s increasingly commercial focus.

Anthropic now competes with OpenAI as well as startups such as Cohere and AI21 Labs, all of which are developing and producing their own text-generating – and in some cases image-generating – AI systems.

“Anthropic has been strongly focused on research for the first year and a half of its existence, but we are convinced of the need for commercialization, which we fully committed to in September [2022]”, reads the pitch deck. “We have developed a strategy for go-to-market and initial product specialization that aligns with our core expertise, brand and where we see adoption in the next 12 months.”

The pitch deck reveals that Alameda Research Ventures, the sister company of Sam Bankman-Fried’s collapsed cryptocurrency startup FTX, was a “silent investor” in Anthropic with “non-voting” shares — responsible for leading Anthropic’s $580 Series B round million. Anthropic expects Alameda’s shares to be sold in bankruptcy proceedings within a few years.

Google is also one of Anthropic’s investors and has pledged $300 million in Anthropic for a 10% stake in the startup. Under the terms of the deal, which was first reported according to the Financial Times, Anthropic agreed to make Google Cloud its “preferred cloud provider” with the companies “co-develop[ing] AI computer systems.”

Other Anthropic backers include Facebook co-founder James McClave and Asana Dustin Moskovitz, former Google CEO Eric Schmidt and founder of Skype engineer Jaan Tallinn.