AI最大的颠覆，可能来自便宜100倍的模型

His answer: context. Levie said the true advantage in an AI-driven economy won't come from having smarter models, but from giving those models access to the right proprietary information — the internal data, customer histories, workflows, decision-making patterns, and institutional knowledge that companies have built over time. "Certainly it will be about how teams and employees use AI agents effectively," he wrote, "but the ultimate force-multiplier will be the context that the agents get." ## Context over prompts The idea is gaining traction across Silicon Valley. [...] Business Insider # AI is turning expertise into a commodity. Box CEO says there's one way companies can stay ahead. Thibault Spirlet 2 min read DATA-USD +5.74% WIN-USD +7.09% TIME-USD -0.76% NOT-USD -0.51% HIGH-USD -4.80% Box CEO Aaron Levie says AI will commoditize expertise, forcing companies to compete on context. As AI agents spread, Levie says proprietary data and institutional know

Russell Brandom The AI boom has been built on a basic assumption: Bigger models are more powerful, and the most powerful models win. Now, the industry is about to learn what happens if that assumption starts to break. Mounting costs have already pressured users to give smaller and cheaper models a second look. This cost-conscious model-shopping is new and it’s unclear how it will affect the industry, but the impact is likely to be significant. One prediction, laid out best by Coinbase co-founder Brian Armstrong, is that it will result in the vast majority of tasks shifting to cheaper models. [...] “[D]emand for intelligence is near infinite, but 80% of workloads will be running on 99% cheaper models within 12-18 months,” Armstrong wrote on X. “20% of workloads will still run on latest gen models where IQ maxing is important.” It’s hard to overstate what a significant shift it will be for the AI industry if Armstrong’s prediction comes true. Before now, most AI companies have compe

When a capability becomes cheap and available to everyone, the instinct is relief. Everyone has it, so no one has an edge, so I can relax. That is precisely wrong. When the tool is universal, the tool stops being the advantage. The advantage moves. It moves to what you point the intelligence at, how fast you move, and whether you have the nerve to aim it at something that matters. Cheap, universal intelligence does not remove the need for bold bets. It raises the stakes on making them the right way. Leave a comment ### The skills that decide who wins When intelligence is a commodity, human skill decides who wins. And the decisive skill is executive leadership's capacity to transform the enterprise. Here is where the value is concentrating. [...] The future just became clear. Frontier AI models and intelligence just became cheap, abundant, and ownable. They were just commoditized. We are riding an accelerating intelligence-to-cost curve, and it bends in your favor. That is great new

The most valuable foundation-model companies may still become some of the largest businesses of this era. But the most important AI companies will not all train models themselves. Many will win by controlling what intelligence alone cannot provide: the context to understand, the workflow to participate and the permission to act. Seed Radar Europe ### Seed Radar Europe #### 405 followers Like Celebrate Support Love Insightful Funny Harsh Jignesh Shah, graphic This feels similar to the evolution of enterprise software. The winners were rarely the ones with the most features, but the ones that became deeply embedded in how organizations operate. AI may follow a similar path. To view or add a comment, sign in ## More articles by Jamie Tyler Gruber [...] The specific benchmark matters less than the direction. Frontier models may remain differentiated, but most businesses do not require the absolute frontier for every task. They need a model that is reliable enough to review a contra

Meanwhile, ADvendio CEO Bernd Bube is advancing AI adoption across media operations, advertising sales workflows and campaign management, while investors such as ElevenX Capital's Anjli Jain have been increasingly identifying opportunities in industry verticals – including education – and backing companies like IvySchool.ai, which helps individuals build the skills needed for an AI-driven economy. As the AI market matures, the next wave of value creation may not come from building larger models alone, but from the companies and leaders that make those models indispensable to everyday business operations. The ventures that successfully connect AI innovation with real-world execution could thus become some of the most sought-after investment opportunities in the years ahead. [...] However, Meta wasn’t buying an AI model or LLM. It was betting on a crucial provider of high-quality training data, a resource that’s becoming scarcer than the models themselves. A year on, Zuckerberg’s play

the moment the market has to decide whether it's valuing a search engine or a compute landlord. And those two businesses, my friends, do not trade at the same multiple. So, stop asking who has the best AI and start asking who owns the pipes underneath it. The model that impresses you today will likely be a commodity within a few years. The power plant, the chip, the chip factory, the data center, those don't get disrupted by next month's software update. So, pay attention to those. So, your truth bomb for today is this. The AI model you talk to is quietly turning into a commodity, but the power, the chips, and the data centers underneath it are the scarce layer for the moment, and whoever owns that layer gets paid long after this round of chatbot wars is forgotten. Join me every day for [...] a third, and the typical business went from running two AI models to nearly five. Now, I want to be straight with you. That data comes from an AI infrastructure vendor, so you know, it's not an in