I sat down with Aaron Levie on the O’Reilly AI Codecon two weeks in the past. Aaron cofounded Field in 2005, and 20 years later, his firm manages content material for about two-thirds of the Fortune 500. Aaron is likely one of the few CEOs of an incumbent enterprise software program firm pondering deeply in public about what AI means for your entire enterprise stack. There are lots of people who’re constructing corporations from the bottom up with AI, others who’re dragging their toes adapting present enterprises to it, after which there’s Aaron. He sits in a sort of Goldilocks zone, enthusiastic however not uncritical, partaking within the laborious work of adapting AI to the enterprise and the enterprise to AI.
The engineering demand paradox
I began out by asking about one thing from Lenny’s Publication that Aaron had retweeted. Regardless of all of the doom rhetoric, TrueUp knowledge exhibits software program engineering job postings are at a three-year excessive. Product supervisor jobs are approach up. AI jobs as a complete are approach up.
The precise knowledge could also be extra equivocal than the TrueUp report suggests. The trustworthy learn of the literature as of spring 2026 (Brynjolfsson et al., Humlum and Vestergaard, BLS Software program Builders, BLS Pc Programmers) is that one thing actual is going on to entry-level software program work, that it’s occurring quicker than most earlier expertise transitions, that it has totally different results relying on which job code you take a look at, and that it isn’t but clear whether or not the web impact on whole software program employment will probably be detrimental, impartial, or finally constructive. Nonetheless, the TrueUp report was a set off for the dialogue that adopted.
Aaron famous that engineers have traditionally been concentrated at tech corporations as a result of the price of a software program undertaking was too excessive to justify anyplace else. But when brokers make an engineer two to 10 occasions extra productive, all of the software program tasks that have been by no means economically viable all of the sudden change into viable. Demand doesn’t shrink. It diffuses throughout your entire economic system. In his tweet, he known as it “Jevons paradox occurring in actual time.” In our dialog, he stated:
“What’s going to occur is your entire world goes to be taking a look at all of the potential software program that they construct. They usually’re going to begin to say, Oh, I can lastly justify going out and doing one of these undertaking the place I couldn’t earlier than.”
Engineers empowered by AI brokers gained’t simply construct software program for IT groups. The full addressable function of the engineer expands from the expertise division to each operate within the enterprise. They’ll be wiring up automation for advertising and marketing, authorized, accounting, and each different company operate.
He’s completely proper. Go searching in any respect the crappy workflows, the crappy processes, the unbelievable overhead of issues that should be easy. You suppose corporations ought to lay off their builders to scale back prices when there’s a lot shitty software program on the market? Actually? There’s a lot that must be improved. He had an incredible line: “Silicon Valley is spooked by its personal expertise.”
Over to me: The rhetoric from the labs about job destruction is actively counterproductive. I used to be speaking not too long ago with somebody in healthcare who described a hospital system making an attempt to fill an enormous gap from diminished Medicare funding. They see AI as a approach to acquire effectivity of their again workplace to allow them to release extra assets for affected person care. And naturally the union is combating it as a result of they’ve been informed AI is a monster that’s going to take their jobs. In case you inform a unique story, one about making the system higher and serving extra individuals extra affordably, that’s one thing individuals can get behind. We’ve to vary the narrative.
Context, not connectivity, is the actual downside
I additionally requested Aaron whether or not protocols like MCP are making context transportable sufficient to erode aggressive moats. He agreed that the trade has broadly converged on openness and interoperability (with some toll cubicles to work via). However getting your methods to speak to one another doesn’t resolve the tougher downside of getting your knowledge structured in order that brokers can truly discover the correct data on the proper second.
“If it’s in 50 totally different methods and it’s not organized in a approach that brokers can readily reap the benefits of, what you’re going to be is on the mercy of how effectively that agent finds precisely the context that it must do its work. And also you’re sort of simply rolling the cube each time you do a workflow.”
He predicts a decade of infrastructure modernization forward, which sounds about proper. At O’Reilly, I maintain operating into this myself. I’ll see a activity that’s excellent for an agent and shortly uncover that the information I want is scattered throughout 4 methods and I’ve to leap via hoops to determine who is aware of the place the information is and how you can get entry. A buddy operating a big (however comparatively new) enterprise that’s turbocharging productiveness and repair supply with brokers informed me not too long ago {that a} huge a part of his crew’s success was doable as a result of they’d spent numerous time getting their knowledge infrastructure so as from the beginning.
IMO, numerous the tales you hear about OpenClaw and different harbingers of the agent future might be deceptive in an enterprise context. They’re doing greenfield setups, largely operating client apps with well-defined interfaces, and even then, it takes weeks to arrange correctly. Now think about agentic frameworks for corporations with hundreds of workers, a whole bunch of legacy apps, and deep wells of proprietary knowledge. A decade of infrastructure modernization is beneficiant. With out assist, many enterprises could have issue making the transition.
Engineering the trade-offs
I introduced up Phillip Carter’s “two computer systems” framing, that we’re now programming a deterministic pc and a probabilistic pc on the identical time. Abilities are a bridge, as a result of they’ve each context for the LLM which may work probabilistically and instruments which might be constructed with deterministic code. Each methods coexist and work in parallel.
Aaron known as the boundary between the 2 computer systems “the trillion-dollar query.” When does a course of cross the edge the place it ought to be locked into repeatable, deterministic code? When ought to it keep adaptive? Mortgage processing must work the identical approach each time. Worker HR queries might be probabilistic. And the irony, as Aaron identified, is that making these trade-offs appropriately requires deep technical understanding. AI makes the sphere extra technical, not much less.
I added that generally this judgment is a person expertise query, generally a price query. You are able to do one thing with an LLM, but it surely is perhaps quite a bit cheaper with canned code. At different occasions, regardless that the LLM prices extra, the flexibleness of a liquid person interface is much better.
That is additionally a locus of creativity. What you convey out of AI is what you convey to it. Steve Jobs wasn’t a coder, however he knew how you can get probably the most out of coders. He would have gone nuts with AI brokers, as a result of he was the essence of style and judgment and setting the bar.
The place startups win
I requested Aaron in regards to the dangers to present enterprises from greenfield AI startups that may simply transfer quicker, reinventing what the incumbents do with an AI native resolution, with out all the bags. He replied:
“If there’s already a considerable quantity of the information for that exact workflow in an present system, and the incumbent is agile sufficient and responsive sufficient, then they’re in a great place to construct both the options or to monetize that set of labor that’s going to be finished….What brokers are actually good at is automating the unstructured areas of labor, the messy, collaborative human-based components of labor, the tax course of, the authorized overview course of, the audit and threat evaluation means of all your contracts and unstructured knowledge. And so in these areas, there’s no incumbent. The one incumbent is probably going skilled companies companies. In order that’s the place I might favor startups.”
Software program startups like Harvey are already taking companies domains and constructing brokers for them. However it’s not simply software program startups. Aaron additionally sees numerous alternative for AI-native regulation companies, accounting companies, and advert companies that may throw away legacy workflow, begin from scratch, and ship two to 5 occasions the output at decrease price could have an enormous benefit.
I did push again with some extent I feel is underappreciated: Current enterprises face an actual threat that the group will attempt to stuff AI into present workflows moderately than asking what the AI-native workflow can be. Individuals are connected to their jobs, their roles, the org chart. We’ve to wrestle with that actually if we’re going to really reinvent what we do.
People get context without cost
One in all Aaron’s factors about brokers is that people carry an infinite quantity of ambient context that brokers lack. You realize what constructing you’re in and who else works there and what they do. You realize the assembly that simply occurred the place a crew modified course on a method that hasn’t been written down but. You will have 20 years of collected area information. All of that’s free context that we’ve by no means needed to formalize. As he put it, “We’ve by no means constructed our enterprise processes in a mannequin the place we assume that there’s a brand new person in that workflow that appeared one second in the past and in underneath 5 seconds, they should get all the data doable to try this activity.”
He advised that a method to consider brokers is as new workers who’re consultants however arrive with zero context and should be absolutely briefed. And the context must be exact, not simply complete. Give an agent an excessive amount of context and it will get confused. Give it too little and it rolls the cube. SKILLS.md and AGENTS.md information are makes an attempt to offer precisely the surgical context an agent wants for a particular course of.
However 99% of information work doesn’t have an AGENTS.md file, he famous. The info is in all places. The context is in all places. So in an present enterprise, it’s important to reengineer workflows from the bottom as much as ship the correct data to brokers on the proper second.
Aaron summed up Field’s strategic pivot in a single sentence: swap the phrase “content material” for “context” and the remainder of the technique stays the identical. Enterprise context lives in contracts, analysis supplies, monetary paperwork. That’s all enterprise content material but it surely isn’t at all times simply obtainable as context. The evolution is making brokers first-class residents alongside individuals as customers of that content material. This very a lot maps to what we’re serious about at O’Reilly too.
