\n<\/p>\n
Of all the debates raging about the potential downsides of AI, there is one worry causing the most hand-wringing among AI enthusiasts in Silicon Valley. Their fear is that the giant AI labs that sell proprietary models are somehow acting like Trojan horses.<\/p>\n
The concern is that, as startups and enterprises use AI models from labs like OpenAI and Anthropic, the labs gain ever-increasing access to those companies\u2019 most sensitive business information. The model makers can then use that knowledge for themselves, potentially becoming competitors to their own customers. Those issuing such warnings range from VCs like Jason Calacanis<\/a> to Palantir CEO Alex Karp.<\/p>\n Now, in a surprising blog post<\/a> published on Monday, Microsoft CEO Satya Nadella has joined this crowd. Nadella warns that AI users (the \u201cbuyers\u201d as he calls them) are paying twice. They knowingly spend for AI token usage but they also, obliviously, hand over valuable data in the process.<\/p>\n \u201cYou essentially pay for intelligence twice, once with money, and again with something even more valuable: the proprietary knowledge you must reveal to make that intelligence useful. The better you want the model to perform, the more of that knowledge you have to feed it!\u201d he writes.<\/p>\n Most dangerously, enterprises are literally teaching the models about the nuances of their businesses, he argues.<\/p>\n \u201cModels learn from \u2018exhaust,\u2019 the prompts people write, the tools agents use, and especially the corrections people make when the model is wrong. Every correction is distilled into institutional know-how,\u201d he writes.<\/p>\n This is \u201cthe kind of knowledge a competitor could never buy,\u201d and yet enterprises are handing it over.<\/p>\n Nadella argues that if AI companies get to freely scrape the internet to train their models, it\u2019s only fair that enterprises get to study \u2014 or \u201cdistill\u201d \u2014 those models in return. \u201cDistillation\u201d is the practice of using a model\u2019s own outputs to learn how it works and to train a new, often cheaper, model based on those insights. In February, Anthropic accused Chinese open source models of sending millions of prompts to Claude<\/a> as a way to improve their own models, and urged the U.S. government crack down on export controls.<\/p>\n Nadella\u2019s point is that model makers can\u2019t have it both ways. It\u2019s hypocritical for them to freely train on the world\u2019s data while restricting others from doing the same to their models.<\/p>\n \u201cWhile the great innovation that comes from model providers having fair use rights to train models on public data is needed, I find it ironic that the status quo is to then turn around and impose restrictive terms on distillation,\u201d Nadella writes.<\/p>\n Nadella is particularly concerned when model makers \u201creserve the right to learn from customer usage and interaction data.\u201d<\/p>\n Nadella\u2019s solution is the kind of thing the CEO of a giant cloud provider would suggest. He wants companies to \u201cretain ownership\u201d of their data, including prompts, feedback, etc. So he\u2019s urging them to build their own \u201cproprietary learning environments\u201d on the cloud (where their data is likely already stored anyway and, conveniently, could mean Microsoft\u2019s cloud, Azure). He also wants companies to build in what he calls \u201corchestration layers\u201d \u2014 essentially, a way to easily switch between AI models from different providers rather than being locked into one. Tools like AI \u201cgateways\u201d that let companies do exactly this have become increasingly popular.<\/p>\n While Nadella never uses the words \u201copen source\u201d as the method for retaining ownership, this is an obvious subtext. Yet, there\u2019s another subtext.<\/p>\n Large companies, many of which still have some of their own data centers in addition to using the cloud, are already moving to open source models installed on their own premises (\u201con-prem,\u201d in industry jargon). Idit Levine, founder and CEO of Solo.io \u2014 which makes networking and security software that helps enterprises manage AI systems \u2014 says she\u2019s seeing exactly this shift play out with her own customers. After experimenting with proprietary model makers, they start asking themselves: \u201cCan I take an open source model and run it on-prem? It will do almost 90% of what the big one\u2019s doing. It will cost way less,\u201d she tells TechCrunch. \u201cThey understand that, and they can control it.\u201d<\/p>\n Solo.io\u2019s technology was selected last year to be the tech powering the Linux Foundation\u2019s Agentgateway project<\/a>. Her company counts enterprises like T-Mobile, ADP, and SAP as customers. She sees companies increasingly installing on-premise open source models and sees it as the next big wave in enterprise AI use.<\/p>\n She\u2019s not alone. Vercel (best known as a platform for building and hosting websites, which has recently added AI model-switching tools) and OpenRouter (a company that helps developers route requests across different AI models) are both seeing a surge in traffic to open source models<\/a>. In fact, open models accounted for 29% of all traffic routed through Vercel\u2019s gateway<\/a> last month.<\/p>\n With the CEO of Microsoft, a company that has invested in both OpenAI and Anthropic, now openly urging enterprises to be wary of using proprietary models, we\u2019ll bet this trend continues to grow. \u201cIn consuming intelligence, you are creating intelligence. And what you create should belong to you,\u201d Nadella writes. <\/p>\n<\/div>\n When you purchase through links in our articles, we may earn a small commission<\/a>. This doesn\u2019t affect our editorial independence.<\/em><\/p>\n