How DeepSeek and Next-generation aI Agents might Erode Value Of Language Models

In the internet transformation, we’re moving from building websites as the company to really developing internet-native business – so, the Airbnb of AI, the Stripe of AI,» he added. «They are not about the design. They have to do with the system and how you make the model helpful for tasks.»

The arrival of DeepSeek’s R1 model last week is viewed by some tech CEOs as a more indicator LLMs are becoming progressively commoditized.

R1 includes some essential aspects that distinguish it from other models on the marketplace. For instance, the model includes a «mixed accuracy» structure that utilizes a combination of full-precision 32-bit floating point (FP32) numbers and low-precision 8-bit drifting point (FP8) numbers.

The latter is faster to procedure but can be less accurate. However, instead of depend on one or the other, DeepSeek uses FP8 for the majority of estimations and changes to FP32 for specific tasks where a greater degree of precision is needed.

«As AI gets more effective and available, we will see its use skyrocket, turning it into a commodity we just can’t get enough of,» Microsoft CEO Satya Nadella composed on social media platform X, in an apparent referral to DeepSeek.

Meanwhile, Matt Calkins, CEO of U.S. software firm Appian, told CNBC that DeepSeek’s success just reveals that AI models are going to end up being more of a product in the future.

«In my viewpoint, we’re going to see a commoditization of AI. Many companies will attain competitive AI, and a lack of differentiation will be bad for high-spending first-movers,» Calkins said through e-mail.