Through my lens

I've worked on productivity tools from a few different angles: capturing information at Evernote, shopping lists at Out of Milk, conversations in education at Edmodo, and construction workflows at PlanGrid. The problems looked different, but I kept coming back to the same questions. How easy is it to get something in? Can you find it again? Does it still make sense when you come back to it?

I really feel that tools built around memory will have the strongest chance of prevailing. By memory, I mean more than a pile of saved notes. I mean remembering the work, the conversations around it, the decisions we made, and why we made them. Then being able to fetch the right pieces when they're useful.

It's getting easier to spin up an app. A notes editor, a task list, even a fairly tailored workflow can be built much faster than before. I don't think another note app can expect to enter the market and take over just by offering a nicer place to type. There needs to be a bigger technical leap in what it helps someone do. Memory and context feel like a big part of that opportunity.

How we got here

Before software, we had notebooks, planners, index cards, and filing cabinets. We put things somewhere so we wouldn't have to keep them all in our heads. The system only worked if we could remember where we'd put them.

Word, Excel, and PowerPoint moved familiar work onto a computer. Documents became easier to edit, copy, and share, but the organizing model was still largely files and folders. You needed to know which file to open and what you were looking for.

Mobile changed when and where we could capture things. A thought on a walk, a receipt, a photo of a whiteboard, something to pick up on the way home. Apps like Evernote, Any.do, and Out of Milk could meet people in those small moments rather than waiting for them to sit at a desk.

That made capture easier. It also gave us a lot more to lose track of. A useful screenshot could disappear into a camera roll just as easily as a document could disappear into a folder.

AI gives us another way into that information. We can describe something we half remember, ask a question across documents, or ask for help turning what we've saved into something useful. For me, that's the more interesting shift. I want the work I've already done to be easier to draw on.

Capturing something is only the beginning

At Evernote, Collect was about reducing the effort of saving useful images. People were already taking photos of documents and receipts. The opportunity was to help get those images out of the camera roll and into a place where they could be cleaned up, searched, and found again.

Evernote Collect interface showing a receipt being reviewed and saved.
Evernote Collect: reviewing and saving a captured receipt. From my work on the product.

That experience still shapes how I think about these tools. If saving something takes too much effort, people won't keep doing it. But easy capture alone doesn't finish the job. The thing you saved has to be useful later.

People also organized their information very differently. Some relied on tags and carefully named notebooks. Others barely organized at all. I don't think everyone should have to become a good librarian to get value from what they've collected.

A tool should help bridge that gap. I might remember that we discussed an idea after a customer interview, but forget the project name or which document it ended up in. Those fragments should be enough to start looking. Once the tool finds something, I want to see where it came from and whether it's still relevant.

The conversation is part of the content

At Edmodo, the context around learning mattered. It needed to be easy to have a conversation alongside the content: a question about an assignment, a clarification from a teacher, or a response that helped someone else understand it.

When the conversation moves somewhere else, you lose part of what makes the material useful. You can find the assignment and still miss the explanation that made it click. In my opinion, even with Google Classroom's position in education, keeping those conversations easy and connected to the content is still something it could do better.

PlanGrid brought a different set of needs into focus. People needed to know which drawing they were looking at, whether it was current, and how an issue related to the work happening onsite. Reliability, access, and version history mattered because other people were acting on that information.

Both experiences made me pay more attention to what surrounds a document. Who was involved? What changed? What did we decide? A file on its own often tells only part of the story.

Where Notion fits

Notion interests me because it gives people room to build their own systems. Pages, databases, and linked information can come together around the way a person or team works. That flexibility is useful, though it can also become another thing to maintain. I've seen how easy it is for organizing the work to become work of its own.

Its May 2025 release brought together AI meeting notes, enterprise search, and research across workspace and connected sources. That's a concrete example of the direction I'm interested in: helping people use information across the places where work happens.

The test for me is what happens after the information is found. Can I understand the answer? Can I check the source? Does it preserve the detail that matters, or flatten a complicated conversation into a tidy summary?

I'd find it much more useful to have a tool bring back a past decision and its reasoning than to generate another page of plausible advice. Especially if it helps me notice that the situation has changed since we made that decision.

Memory needs to earn trust

Storing more information doesn't automatically make a tool smarter. Old plans get replaced. A passing thought isn't a commitment. Something I said in one project may not apply to another.

A memory system has to handle those differences. It needs to preserve sources, understand which information is current, and respect who can see what. I should be able to correct what it remembers or tell it to forget something. Otherwise, an early misunderstanding can keep turning up in later work.

There's also a practical retrieval problem. Finding a paragraph with similar words isn't always enough. The useful answer might depend on a discussion, a drawing, and a decision recorded weeks apart. Connecting those pieces without losing their meaning is hard. That's where I think there's room for real technical progress.

I want the tool to make its reasoning easy to inspect. Show me the conversation it used. Let me open the original. If it can't find enough to support an answer, I'd rather know that than have it fill in the gaps.

Building an app is easier. Understanding my work is still hard.

As it gets easier to build software, I think the bar for a new productivity tool goes up. A polished editor or a familiar task board can still be useful, but those features alone give people less reason to move their work.

What would make me move is a tool that helps me do something I couldn't easily do before. It could pull together the history of a project, reconnect a piece of research with a problem I'm solving now, or help me build a small app around information it already understands.

For example, imagine asking for a project review tool that uses the team's existing decisions, open questions, and customer feedback. I shouldn't have to start by copying everything into a new database and explaining the project from scratch. The tool could help assemble what I need, show me what it used, and let me adjust it.

That's the kind of superpower I see an opportunity for. Being able to spin up an interface is useful. Being able to give it the right context makes it much more valuable.

It also changes how I think about interoperability. If the useful history is spread across notes, messages, files, and calendars, a tool needs ways to work across them. Exporting the text alone won't preserve all the relationships that made it useful. I'd want to keep the sources, connections, and history too.

More output isn't the same as getting further

AI makes it easy to get to something that looks finished. A clean summary or a long document can feel like progress, but the real test is whether someone can use it to move the work forward. If they have to reconstruct the context or correct the details first, some of that work has just been passed along.

Better memory could help ground that output, but it won't remove the need for judgment. A tool can retrieve the wrong thing or apply an old decision to a new situation. The person using it needs a way to catch that before it spreads.

I'd judge these tools by how much they help me move the work forward. Did I spend less time finding the background? Did I avoid repeating a mistake? Could I make a better decision? Those are more useful signals to me than how many notes or drafts it generated.

What I'm looking for next

The thread through all of this, from Evernote to the tools being built now, is making what we know easier to use. Capture matters. So do conversations, retrieval, and the ability to act on what comes back.

I think memory will be a deciding advantage for productivity tools. The ones I'd bet on will remember enough of the right things to become more helpful over time, while still letting people inspect, correct, and move that information.

I don't need another place to put a note. I want help connecting it to what I'm doing next.

Further reading

Best,

Craig Aucutt