Why does the anthropier bet on his future on the factors of artificial intelligence coding

Hello, welcome to you Decoder! I am Hayden Field, Amnesty International correspondent at freedom And the host of the guest of the episode Thursday. I will be hacked in Nelay for a few other episodes, and I am excited to continue to dive into good, bad, and doubtful in making artificial intelligence.

Today, I am talking to David Hershey, who leads the Antarbur Applied Intelligent Intelligence Team. David works with startups to help them know the best way to apply anthropologist, as well as test new AI models to understand their borders.

I wanted to get David because the Antarbur issued a New you have a model Claude Sonit 4.5 is called earlier this week, and it makes the waves. (As a reference, Claude is Anthropor what Chatgpt is to Openai.)

The new model, Sonnet 4.5, is described as a large penetration of independent artificial intelligence, especially for coding purposes. These types of artificial intelligence products can be given, in theory, complicated tasks, then explode and complete for long hours or even several days. Anthropor says that this particular model can work for up to 30 hours in a row without any human intervention – all while working on an individual mission, such as building an application from zero point.

During the past year or so, companies such as Anthropor, Microsoft and Openai were, and more promised that this technology is an agent that will be the next stage of artificial intelligence, the next thing full of noise full of noise that comes after chat stations for general purposes. They say it can really open the potential of artificial intelligence, and it is true that they have taken some steps.

But as we have seen so far, the agents are not there yet, and they have ways to go. In fact, most of us do not send agents on the Internet to provide our bids, and certainly do not give them tasks that may take 12, 24, or even 30 hours of independent work without any human hand. At least, not yet.

At the same time, many companies view the agents as the penetration that is supposed to open huge production gains from artificial intelligence models, including the opportunity to use it to replace or increase human workers.

So I wanted to sit with David, who spends a lot of time to test what situations like Claude Sonnet 4.5 that you cannot do, ask him where we are on this promise of artificial intelligence agents. I wanted to talk about what these types of products are good from the consumer’s point of view, further than programming purposes, and also what the path looks forward with the progress of agents.

If you want to read more about what we talked about in this episode, check the links below:

_{Questions or comments about this episode? Hit us in Decoder@theverger.com. We really read every email!}