About six weeks ago, I despatched an e-mail to Satya Nadella complaining concerning the monolithic winner-takes-all structure that Silicon Valley appears to ascertain for AI, contrasting it with “the structure of participation” that had pushed earlier know-how revolutions, most notably the web and open supply software program. I suspected that Satya may be sympathetic due to previous conversations we’d had when his e-book Hit Refresh was revealed in 2017.
I made the case that we want an structure for the AI business that permits cooperating AIs, that isn’t a winner-takes-all market, and that doesn’t make present corporations in each business merely the colonial domains of extractive AI conquerors, which appears to be the Silicon Valley imaginative and prescient.
Little did I do know that Microsoft already had one thing within the works that was a strong demonstration of what I hoped for. It’s referred to as NLWeb (Pure Language Net), and it’s being introduced as we speak. Satya supplied O’Reilly the prospect to be a part of the rollout, and we jumped at it.
My concepts are rooted in a notion about how know-how markets evolve. We have now lived by means of three eras in computing. Every started with distributed innovation, went by means of a interval of fierce competitors, and ended with monopolistic gatekeepers. Within the first age (mainframes), it was IBM, within the second (PCs) Microsoft, and within the third (web and cellular) the oligopoly of Google, Amazon, Meta, and Apple.
The error that everybody makes is a rush to crown the brand new monopolist at the beginning of what’s primarily a large open area firstly of a brand new disruptive market. When the non-public laptop challenged IBM’s hardware-based monopoly, the race was to turn out to be the dominant private laptop {hardware} firm. Microsoft received as a result of it realized that software program, not {hardware}, was the brand new supply of aggressive benefit.
The story repeated itself firstly of the web period. Marc Andreessen’s Netscape sought to exchange Microsoft as a dominant software program platform, aside from the web slightly than the PC. AOL realized that content material and group, not software program, was going to be a supply of aggressive benefit on the web, however they made the identical mistake of assuming the tip sport of consolidated monopoly slightly than embracing the early stage of distributed innovation.
So right here we’re firstly of the fourth age, the age of AI, and as soon as once more, everyone seems to be dashing to crown the brand new king. A lot of the chatter is whether or not OpenAI or one in every of its rivals would be the subsequent Google, when it seems to be to me that they’re extra probably the subsequent Netscape or the subsequent AOL. DeepSeek has thrown a bomb into the coronation parade, however we haven’t but totally realized the depth of the reset, or conceptualized what comes subsequent. That’s sometimes discovered by means of a interval of distributed innovation.
The time period “the structure of participation” initially got here to me as a proof of why Unix had succeeded as a collaborative mission regardless of its proprietary license whereas different initiatives failed regardless of having open supply licenses. Unix was designed as a small working system kernel supporting layers of utilities and purposes that might come from anybody, so long as they adopted the identical guidelines. Complicated behaviors may very well be assembled by passing data between small packages utilizing normal information codecs. It was a protocol-centric view of how advanced software program techniques ought to be constructed, and the way they might evolve collaboratively. The web was additionally developed as the same distributed, protocol-based system.
That idea ran by means of my net advocacy within the early ’90s, open supply advocacy within the late ’90s, and Net 2.0 within the aughts. Participatory markets are progressive markets; prematurely consolidated markets, not a lot. The limitations to entry within the early PC market have been very low, entrepreneurship excessive. Ditto for the Net, ditto for open supply software program and for Net 2.0. For late Silicon Valley, fixated on untimely monopolization by way of “blitzscaling” (assume Uber, Lyft, and WeWork as examples, and now OpenAI and Anthropic), not a lot. It’s turn out to be a type of central planning. A small cadre of deep-pocketed buyers choose the winners early on and attempt to drown out competitors with huge quantities of capital slightly than permitting the experimentation and competitors that enables for the invention of true product-market match.
And I don’t assume we’ve that product-market match for AI but. Product-market match isn’t simply getting a number of customers. It’s additionally discovering enterprise fashions that pay the prices of these companies, and that create worth for greater than the centralized platform. As Invoice Gates famously informed Chamath Palihapitiya when he was operating the nascent (and finally failed) Fb developer platform, “This isn’t a platform. A platform is when the financial worth of everyone that makes use of it, exceeds the worth of the corporate that creates it. Then it’s a platform.”
To be clear, that isn’t simply worth to finish customers. It’s worth to builders and entrepreneurs. And meaning the chance to revenue from their improvements, to not have that worth instantly harvested by a dominant gatekeeper. We might count on that within the later, enshittified stage of the market, when it’s ripe for disruption, however we don’t need to see it early on!
That’s why I’ve been rooting for one thing completely different. A world the place specialised content material suppliers can construct AI interfaces to their very own content material slightly than having it sucked up by AI mannequin builders who provide up companies based mostly on it to their very own customers.
Way back to our 2017 dialog, Satya had referred to AI as “the third runtime.” That’s, Home windows was the primary mass market software program runtime; the online was the second. (Arguably, cellular was the third however for this function, let’s take into account it an extension of the second.) What may that third runtime appear like? An enormous centralized platform? I hope not. Firms resembling Salesforce and Bret Taylor’s Sierra are betting on brokers which can be frontends to corporations, their companies, and their enterprise processes, in the identical means that their web sites or cellular apps are as we speak. Others are betting on client-side brokers that may entry distant websites, however typically by calling APIs and even performing the equal of display screen scraping.
Anthropic’s Mannequin Context Protocol, an open normal for connecting AI brokers and assistants to information sources, was step one towards a protocol-centric imaginative and prescient of cooperating AIs. It has generated loads of well-deserved enthusiasm. Google’s A2A is a futuristic imaginative and prescient of how AI brokers may cooperate. That is all going to take years to get proper.
However all corporations want at the least a begin on an AI frontend as we speak. There’s a wonderful line from C. S. Lewis’s novel Until We Have Faces: “We can’t see the gods nose to nose till we’ve faces.” Proper now, some corporations are capable of provide an AI face to their customers, however most don’t. NLWeb is an opportunity for each firm to have an AI interface (or just “face”) for not simply their human customers however any bot that chooses to go to.
NLWeb is totally appropriate with MCP however provides one thing a lot less complicated, and for present web sites, rather more instantly actionable: a easy mechanism so as to add AI search and different companies to an present net frontend. We put collectively our demo AI search frontend for O’Reilly in a number of days. We’ll be rolling it out to the general public quickly.
