Of all the ongoing debate about the potential downsides of AI, there’s one concern that’s causing the most concern among AI enthusiasts in Silicon Valley. Their fear is that giant AI labs selling proprietary models act as Trojan horses of sorts.
The concern is that as startups and enterprises use AI models from labs like OpenAI and Anthropic, the labs gain increasing access to those companies’ most sensitive business data. Model makers can then use this knowledge for themselves, potentially competing with their customers. It varies among those issuing such warnings VC as Jason Calacanis To Palantir CEO Alex Karp.
Now, surprisingly blog post Published on Sunday, Microsoft CEO Satya Nadella joined the crowd. Nadella warns that AI users (“buyers,” as he calls them) are paying twice. They knowingly spend money to use AI features, but they also hand over valuable information in the process.
“You’re essentially paying for intelligence twice, once with money and something even more valuable: the specific knowledge you have to uncover for that intelligence to be useful. The better you want the model to perform, the more knowledge you need to feed it!” he writes.
Most dangerously, businesses are literally training models about the nuances of their business.
“Models learn from the ‘exhaust,’ the proposals people write, the tools agents use, and especially from the corrections people make when the model is wrong. Each correction is distilled into institutional know-how,” he writes.
It’s “the kind of knowledge that a competitor can never get,” and yet businesses transfer it.
Nadella argues that if AI companies freely use the internet to train their models, it’s only fair that businesses learn or “distill” those models in return. “Distillation” is the practice of learning how a model works and using its results to develop a new, often cheaper model based on those insights. In February, Anthropic accused China of open source models Cloda sends millions of requests As a way to improve its models, it called on the US government to tighten export controls.
Nadella’s point is that model makers can’t have it both ways. It is hypocritical for them to freely train on the world’s data while restricting others from doing the same with their models.
“While much innovation is needed from model providers with fair use rights to train models on public data, I find it ironic that the status quo would then turn around and impose restrictive conditions on distillation,” Nadella writes.
Nadella is particularly concerned when model makers “reserve the right to learn from customer usage and interaction data.”
Nadella’s solution is something the CEO of a giant cloud provider would suggest. It wants companies to “retain ownership” of their data, including tips, reviews, and more. So he urges them to build their own “proprietary learning environments” in the cloud (where their data is already stored, which could conveniently mean Microsoft’s cloud, Azure). He also wants companies to build what he calls “orchestration layers” — essentially a way to easily switch between different providers’ AI models, rather than being locked into one. Tools like AI “gateways” that allow companies to do just that are becoming increasingly popular.
Although Nadella never used the words “open source” to retain ownership, it is open source. However, there is another subtext.
Large companies, many of which still have their own data centers in addition to using the cloud, are now moving to open source models installed on their premises (“on-prem”, in industry jargon). Idit Levine, founder and CEO of Solo.io, which makes networking and security software that helps businesses manage AI systems, says she’s seeing exactly this shift happen with her clients. After testing with proprietary model makers, they start asking themselves, “Can I take an open source model and run it in-house? It’ll do almost 90% of what the big one does. It’ll cost less,” he tells TechCrunch. “They understand it and they can control it.”
Solo.io technology was chosen as the technology that powers it last year The Linux Foundation’s Agentgateway project. His company counts enterprises such as T-Mobile, ADP and SAP as clients. He sees more and more companies deploying native open source models and sees this as the next big wave in enterprise AI usage.
He is not alone. Vercel (best known as a platform for building and hosting websites, recently added AI model replacement tools) and OpenRouter (a company that helps developers route requests to various AI models) are both seeing traffic growth. open source models. In fact, open models accounted for 29% of all routed traffic Through the door of Versel last month.
With the CEO of Microsoft investing in both OpenAI and Anthropic now publicly urging enterprises to avoid using proprietary models, we’re sure this trend will continue to grow. “By consuming intelligence, you create intelligence. And what you create should be yours,” writes Nadella.
When you purchase through links in our articles, we may earn a small commission. This does not affect our editorial independence.
