This post was originally published on this site.
Anthropic, the maker of Claude AI and backed by Amazon and Google, is not planning to get into hardware or consumer entertainment but will remain focused on developing generalist foundation models for enterprise use cases.
During a fireside chat at the Human[X] conference on Monday (March 10), Anthropic Chief Product Officer Mike Krieger said the startup will channel its resources into enabling use cases for all industries through its Claude family of artificial intelligence (AI) models, essentially balancing research breakthroughs with practical product development.
“We want to help people get work done, whether it’s code, whether it’s knowledge work, etc.,” he said. “And then you can then imagine different manifestations of that” in applications for the consumer, small business and all the way to large corporations and the C-suite.
That’s different from OpenAI, which is not only working on foundation AI models but also building data centers and partnering with former Apple design chief Jony Ive on a potential consumer device.
Read more: Report: Altman May Have Found AI Cash in Powell Jobs’ VC Fund
No Hardware in Anthropic’s Future
Krieger said Anthropic’s approach is exemplified through its product tiers: the consumer-oriented Claude.ai with its $20 per month Pro subscription, solutions for mid-market companies that are often driven by CEOs or CTOs, and enterprise offerings.
Krieger further noted that even Claude’s consumer use cases tend to be practical rather than entertainment-focused, citing examples of parents using Claude to manage their children’s school calendars.
The startup is less interested in making Claude a “chatty … billion-user service” used for consumer entertainment, and more focused on pragmatic and professional use cases, Krieger said. But he acknowledged that there was still “dramatic underbuilding” in consumer AI applications.
Anthropic’s recent partnership with Amazon for Alexa+ is one such example of letting a partner interact with the user. Claude is one of several models powering the new Alexa by providing frontier models and AI expertise; Amazon contributes a global distribution channel through Alexa’s 600 million devices as well as integration capabilities.
“When you think about what we can bring to the table, it’s frontier models and the know-how about how to make those models work really well for really complex use cases,” Krieger said.
However, the startup itself isn’t focused on introducing its own hardware, according to Krieger.
Read more: Amazon Unveils an Agentic, Smarter and More Capable Alexa+
AI to Replace Engineers?
What Anthropic is focused on is improving multimodal capabilities as a key area for further development, he said, noting that Claude’s vision capabilities need enhancement to fully leverage the model’s existing intelligence. “I think our models don’t see as well,” Krieger admitted, noting that improving this capability would unlock new applications.
Anthropic, at its core, is a research-driven company. Krieger described the “dance” of aligning product development with research breakthroughs. This requires operating quickly while maintaining flexibility, as some capabilities arrive sooner than expected while others take longer to materialize.
He noted that decisions about what data should be included in model retraining need to be made well in advance of product launches, requiring a clear vision of where the company wants to be six months ahead.
Asked whether AI like Anthropic’s will replace engineers, Krieger said AI might affect job prospects but noted that Anthropic is still hiring engineers across its desktop, mobile and web products. He believes that AI tools let designers and product managers prototype more effectively.
“In the long run, though, I do think that the way that will manifest is you just need [fewer] engineers,” Krieger explained.
As for what comes after current AI advancements, Krieger mentioned the next frontier will be autonomy and “moonshots” focused on novel discoveries that are currently only limited by the “speed of exploration of ideas.”