What would you do if you were told the problem you are working on will NEVER make money?
"Data portability doesn't make money."
I heard this for years - from market leaders, from VCs, from people I respected.
"This is a regulatory problem, not a technical one." "This is a feature, not a product." "People don't pay for idealistic things."
<<Back story>>
In 2019, my team and I went deep into Self-Sovereign Identity: wrote research papers, ran experiments, and found ourselves at the intersection of data, identity, and web3. Right at that intersection lay data portability and sovereignty: the ability to own your data and take it anywhere on the internet.
This idea wouldn't let us go. No matter how hard we tried to move on, we kept coming back.
We pivoted through doubt, market pressure, and plenty of moments where the easier path was to drop the idea entirely. We thought about leaving this vision and building something else multiple times, but just couldn't. It felt like this problem mattered.
We went through multiple iterations: data plugs, smart profiles, open context layer. But the core conviction never changed: your digital data should be like a USB stick, pluggable into whichever platform you connect to, sharing only what you allow.
Because we believed the infrastructure question — who controls your context, and can you move it freely? — would eventually become impossible to ignore.
That moment is now. AI systems are eating every workflow, and context lock-in is the new data lock-in.
So we built AI Context Flow: the open context layer for AI systems.
Move from Claude to Gemini to Codex to OpenClaw to LMStudio to TypingMind to literally any website on the internet. Your context stays with you, reusable on every platform, and grows with you as you browse through the internet or from within AI Chats.
Proud of the path. Even prouder of staying on it. With a slightly deranged and obsessed team.
What's a conviction you held onto that the market eventually caught up to?
Replies
That you don't need a traditional CS degree to build robust infrastructure. Spent years being told you can't build 'real' software without writing raw syntax line-by-line, but the shift toward natural language logic and autonomous backends proved that clear thinking matters way more than fighting semicolons.
@bellamynina698 times sure are changing. I think we will see a similar shift with "you can't vibe code production level apps" in a couple years.
This resonates.
I don't know yet whether the market will fully catch up, but one conviction I've held onto is that interview preparation should feel like a real conversation, not a questionnaire.
For a long time, most tools focused on generating questions and answers.
I've always felt the harder problem was simulating recruiter behavior: interruptions, skepticism, follow-ups, and the emotional side of evaluation.
Maybe I'm early. Maybe I'm wrong.
But some ideas are difficult to let go of once you've seen the problem clearly.
One conviction I've held onto is that learners and professionals should own their development records rather than having them trapped inside institutional systems. It wasn't always viewed as a business opportunity, but portability and ownership seem increasingly important in an AI-driven world.
@hira_siddiqui1 That's great to hear. I've been exploring similar questions around learner assessment and professional development records. It feels like we're moving toward a future where people expect to carry their evidence, competencies, and context with them rather than leaving them behind when they change institutions or employers.
Honestly, I'd probably keep going but get more strategic about the packaging. "Data portability doesn't make money" is almost always a timing problem disguised as a product problem. The conviction matters, but so does surviving long enough to be right. What I find interesting about your story is the pivot from SSI to AI context portability, that's not a pivot away from the vision, it's the same idea just finally landing in the right moment.
Context lock-in in AI systems is the exact same fight as data lock-in, just with a much bigger addressable problem now. What does the commercial model look like at this point? Still figuring it out or is there a path that's starting to work?
@hira_siddiqui1 It's great to hear that, as long as the model is doing good then you can make money gradually
I'd keep building anyway — but I'd change what I was optimizing for. The honest answer is that 'will never make money' is almost always a distribution problem, not a product problem. Nav-vera operates in a space where the problem is very real — retail operators lose sales every day because customers can't find products. If that were declared 'never profitable' I'd find a different model: licensing the tech, white-labeling it, or going open source for adoption. The problem worth solving usually finds a business model eventually.
Same conviction, different vertical. We heard “broker import is a feature” for years—now context lock-in in wealth-tech means warehousing itemized ledgers just to run AI. We stayed on portable, bounded context: parse at the edge (MIT OSS importer), compress to a fixed aggregate, stateless inference—your USB stick, but for portfolio signal, not chat logs. Proud you kept going on context portability; we’re doing the regulated slice. 🙌
I’d keep building, but I’d change what I was selling. The problem is rarely the problem — it’s the packaging. ‘Relationship intelligence’ might not make money. But ‘the tool couples use before therapy’ might.
@dani_mashael exactly, like for us it was not "data portability", it was "stop repeating yourself to AI in every new conversation and save 5+ hours a week".
I'm more interested to read the replies here rather than to add anything further.
Timing really is everything. The same idea that sounds impossible in 2019 becomes obvious in 2025. Glad you weathered the doubt.
I dont believe any type of problem can make money. Maybe the solution is too early or not the right one. But still you can make money out of any problem, you just treat it a/b testing until you find what buyers re willing to pay. In most cases you'll have to interview enough potential buyers to find out what exactly is the pain in their work and how to market and price your solution.
@hira_siddiqui1 You never know how many are whiling to pay until you actually get the money. For my experience many will say interesting, many will like, comment or whatever in social, some will try it also, but once you have the right problem solved they ask you for the price. But even then you dont know at scale if many will pay. The only way to know how many people are willing to pay is to build and sell something already existing in the market so you use them as reference, like is says tools etc. But then you have other problems coming
@vasileiost you are so right. Either you can be another product solving the same problem in a market or you can create the market.
The problem with creating your own market though is that you either go to the moon or go to zero - which is what a startup is anyways.
@hira_siddiqui1 Sure, but for this it all comes to timing, patience/time available and capital. Some times it's a slow painful process. Not for everyone and not always there is a market to be created/build.