For thirty years we've been told the internet would level the playing field.
It hasn't. Not really. Consumers still get taken advantage of on car deals. Small business owners still sign vendor contracts they don't fully understand. Patients still pay medical bills they suspect are wrong. Independent operators still run their businesses half-blind against competitors with real analytics.
The strange part is that most of the information people needed to avoid these outcomes has been technically available for years. So why hasn't the playing field leveled?
Because "information availability" isn't a single thing. It's a five-stage arc, and each stage has a hard ceiling on what it can actually do for the person holding the weaker position in a transaction.
The Foundation: Asymmetry Is the Default
The economist George Akerlof won the 2001 Nobel Prize in part for showing why this matters. His 1970 paper "The Market for Lemons" — rejected by three major journals before the Quarterly Journal of Economics finally accepted it on the fourth attempt — demonstrated that when one side of a transaction has more information than the other, markets don't just become unfair. They can collapse entirely. The good cars get pulled from the market because buyers, unable to distinguish them from lemons, won't pay what they're worth. Only lemons remain.
The insight wasn't just about cars. Akerlof extended it to insurance, credit markets, labor markets, and healthcare. The economy is full of these asymmetries, and he showed that the partial remedies we've developed — warranties, certifications, intermediaries, regulators — are expensive, imperfect workarounds for a structural problem the market can't solve on its own.
Every wave of information technology since has been a new attempt to close those asymmetries. None has fully succeeded. Each has gotten us one stage closer.
Stage 1 — Information Not Available
Pre-printing-press. Information existed, but it lived inside institutions and specialists. In medieval Europe, a single handwritten manuscript could take a scribe a year or more to produce and cost a fortune. Literacy was confined to clergy, nobility, and a thin layer of merchants — historians estimate roughly 30% of European adults were literate when Gutenberg's press arrived around 1440. If you wanted to understand your lease, you hired a lawyer. If you wanted to evaluate a supplier, you found someone in the industry. The only path to useful knowledge was a relationship with someone who had it.
Gutenberg's press was the first meaningful crack in this. Within fifty years, an estimated ten to twelve million volumes were in print across Europe, up from around thirty thousand handwritten manuscripts before. Literacy climbed — reaching roughly 47% in England by the late 1600s, and 62% by 1800. Movements that had been confined to cloisters — the Reformation, the Scientific Revolution, the Enlightenment — were suddenly possible because information could travel faster than the institutions trying to control it. The press didn't eliminate the gatekeepers. But it put the first real pressure on them.
Stage 2 — Information Available but Difficult to Access
Early internet. The information started showing up online, but the cost of getting to it was still high. You needed to know it existed, know where to look, and be able to parse an unfamiliar system. Government agencies posted PDFs. Industry publications went digital. Libraries scanned collections. But you had to bring your own search strategy, and most people didn't have one.
The gatekeepers held most of their leverage. The professions that had thrived on Stage 1 — law, medicine, finance, real estate — kept their power largely intact. They just had to deal with slightly more informed clients.
Stage 3 — Information Available and Accessible
The search era. The search bar collapsed the access cost. Anyone could find almost anything. Health information, property comps, salary data, legal guidance, business reviews, regulatory filings. The raw material was suddenly free. This was the moment when consumer and small business empowerment felt inevitable.
And yet outcomes barely moved. Because raw information isn't the same thing as usable intelligence. A thousand search results about mortgages don't help a first-time buyer at 10pm staring at a loan estimate, trying to decide whether the rate is good or whether they're being gamed. A directory of vendor reviews doesn't help a gym owner evaluate whether the specific contract in front of them is fair.
Stage 4 — Information Contextualized but Out of Context
The app era. Products got smarter. Personal finance tools categorized your spending. Real estate apps showed comps for your search. Credit tools personalized your score. Small business software organized your finances. Sales tools organized your pipeline. The information was now structured around the user, not just presented as a pile.
This is the stage we've been stuck in for the last decade. And it has a hard ceiling.
Because "contextualized around you as a user" still isn't the same as "contextualized around the specific decision you're facing right now." A market dashboard can tell you the average for a neighborhood. It can't tell you whether the specific offer you're drafting tonight is reasonable given this specific house, this specific seller's motivation, this specific week's signal. A finance tool can show you your margin by category. It can't tell you whether the new vendor contract on your desk is worse than what comparable businesses accept. A credit product can show you your score. It cannot walk into a specific dispute with you and tell you which procedural lever will actually get the bureau to respond.
The Stage 4 products that tried to serve consumers directly ran into a second problem beyond the technology ceiling: the economics of distribution. Acquiring a user who only needs the tool once a decade — for a home purchase, a divorce, a major medical decision — is brutally expensive. The value of better judgment in that moment is high, but the frequency is low, so most of these products had to find revenue by referring users to the very institutions the tools were supposed to counterbalance. The empowerment promise got quietly monetized away. Even well-intentioned public-sector efforts — with no revenue pressure, no conflicts, genuine mission alignment — produced well-organized libraries of generic guidance that still cannot tell any individual what to do about the specific document in front of them.
The ceiling of Stage 4 is the ceiling of the entire consumer and small business empowerment decade. It's why a decade of well-funded attempts failed to change outcomes for the people they promised to help. Not because demand was low. Not because willingness to pay was low at the moment it mattered. Because the technology could only get to "here's information organized around you" — and the business model had to monetize somewhere the user wasn't actually willing to pay.
Stage 5 — Information in Context
Now. For the first time, a product can take your specific situation — this loan estimate, this medical bill, this vendor contract, this member drifting in your gym, this offer on your desk — and produce an opinion about that exact case, at the moment you need it.
Not guidance in general. Not data organized around you. Judgment about the specific thing in front of you, right now.
The economics of this shift are easy to miss if you're not watching closely. Stanford's 2025 AI Index reported that the cost of running inference at GPT-3.5-equivalent performance fell from roughly $20 per million tokens in November 2022 to $0.07 per million tokens by October 2024 — a drop of roughly 280x in under two years. Epoch AI found that price declines across performance levels ranged from about 9x per year to more than 900x per year, with the fastest declines beginning after January 2024 and accelerating. Andreessen Horowitz has documented roughly a 10x annual decline at fixed performance levels. These are among the fastest unit-economics shifts in the history of commercial technology.
What that means in practical terms: judgment that previously required a senior expert charging hundreds of dollars an hour to produce — a read on your specific situation, personalized, reasoned, grounded in your actual context — can now be embedded in a product that costs less than a streaming subscription. That's not an incremental improvement. It's a category change.
I want to be clear about what cheaper inference does and doesn't guarantee. It doesn't automatically translate to better judgment. A model running on a mortgage document for pennies still has to get the analysis right, and hallucination rates on specific, reasoning-heavy tasks remain high enough that production-ready Stage 5 products require real engineering — retrieval systems, domain-specific evaluation, careful interface design that knows when to say "I don't know." The cost collapse is necessary but not sufficient. The companies that build Stage 5 products well will be the ones that respect that gap between raw capability and reliable output, and invest accordingly.
What Stage 5 Actually Looks Like
The first-time car buyer doesn't need another pricing page. They need a product that reads the specific offer on the desk and says: the price is above market for this trim this week, the financing is above what your credit profile qualifies for, the add-ons are where most of the margin lives and here's which ones are genuinely worth it for your situation. Here's what to say.
The gym owner doesn't need another dashboard. They need a product that looks at the specific members drifting in their specific gym and says: these three are 80% likely to cancel in the next 30 days, here's what the research says actually brings them back, here's which coach should have that conversation, before class tomorrow.
The small business owner doesn't need a library of contract templates. They need a product that reads the specific MSA the vendor just sent over and flags which clauses are worse than what comparable businesses accept, and drafts the counter-language.
Stage 5 is what the technology finally allows. The institutions that benefited from Stage 4's limitations — the ones whose margin depended on the fact that the other side couldn't parse the document in front of them — are about to have a decade of accumulated asymmetry challenged by products that were simply impossible to build before 2023.
What This Doesn't Do
I want to be honest about what this doesn't do. The institutions will use the same tools. The dealer still knows the floor price better than any product does. The hospital billing department still has more lawyers than you. They'll use Stage 5 capabilities to build contracts and processes specifically optimized against the kinds of pushback Stage 5 tools enable. The arms race is underway and it's not going to stop.
Not every asymmetry is an arms race. Some of what costs ordinary people and small businesses isn't the institution's superior weaponry — it's the usability gap. The cognitive cost of parsing the thing in front of you, for a decision you make once every ten years, is what makes outsourcing feel mandatory.
That specific asymmetry — rooted in usability, not in fundamental power — is the one where the balance can actually shift. Not parity. But enough leverage to change the outcome on the decisions that matter most.
The Question
That's what A8C Ventures is built around. Not another Stage 4 product with a better interface. Stage 5 products for the markets where the usability gap has been widest, the failure has been most visible, and the people and businesses on the wrong side of it have been underserved the longest.
Consumers buying cars and homes. Patients disputing bills. Small businesses negotiating with vendors and insurers. Independent operators running recurring-revenue businesses half-blind. These aren't separate problems with separate solutions. They're the same problem at the same moment in the same arc.
The question isn't whether Stage 5 products will get built. They will. The question is whether they'll be built by the people who understand what the previous stages actually failed at — or by the people who just want to slap AI on a Stage 4 business model and call it new.
The tools are finally here. The question is whether anyone builds them right.
A8C Ventures is a venture studio building Stage 5 products for high free-cash-flow service businesses underserved by the previous era of software. If you're building where the usability gap has been widest, we should talk.