AGENDA: A Personal Information Manager (PIM)

Basic | The Personal Computing Platform

We build products, tools, and standards to make computers more personal.

Unbundling Tools for Thought

Disparate thoughts on personal information management after traversing the hype cycle.

All-Purpose Machines

APM is a home for a collection of R&D efforts into the future of software, HCI, and personal workstations. While computing has a long way to go before it fullfils its potential, these experiments are designed to embrace and enhance existing, pervasive, well-understood technologies rather than reinvent the wheel. Research projects include:

Things 3 From an OmniFocus and TaskPaper User

I spend a lot of time with my task manager. My day job is a little bit of technical development and a little bit of project management. The project management piece is a capital “P.M.” I manage a dozen large projects spread over several years. Each project has between one and twenty different people on the team. This review is written from that perspective. I’m not a productivity blogger or a guy that just likes to try out software.

The Panoptic Sort: A Political Economy of Personal Information

A Political Economy of Personal Information

The Interfaces With Which We Think

The concepts in modern operating systems — apps, windows, desktops, notifications, and so on — have so permeated our understanding of personal computing, it’s hard to imagine anything else, let alone believe there could be anything better.

What have note-taking PKMs accomplished, really?

As somebody who writes a lot, I've always been interested in personal knowledge management systems and note-taking apps. But what do these frameworks and m…

~𝔐𝔦𝔩𝔰𝔢𝔱 on Twitter / X

To catch up on the current state of @urbit, check out Hawk: a sort of "namespace computer", rethinking how you manage data on your personal cloud. Over the next 15 short videos, you'll see how it unlocks a fully programmable environment you control. pic.twitter.com/gcB4l1iY1Z— ~𝔐𝔦𝔩𝔰𝔢𝔱 (@0xsyru) December 21, 2024

Andrej Karpathy on Twitter / X

The race for LLM "cognitive core" - a few billion param model that maximally sacrifices encyclopedic knowledge for capability. It lives always-on and by default on every computer as the kernel of LLM personal computing.Its features are slowly crystalizing:- Natively multimodal… https://t.co/2jsVevkTSJ— Andrej Karpathy (@karpathy) June 27, 2025

Information scraps: How and why information eludes our personal information management tools

In this article we investigate information scraps —personal information where content has been scribbled on Post-it notes, scrawled on the corners of sheets of paper, stuck in our pockets, sent in email messages to ourselves, and stashed in miscellaneous digital text files. Information scraps encode information ranging from ideas and sketches to notes, reminders, shipment tracking numbers, driving directions, and even poetry. Although information scraps are ubiquitous, we have much still to learn about these loose forms of information practice. Why do we keep information scraps outside of our traditional PIM applications? What role do information scraps play in our overall information practice? How might PIM applications be better designed to accommodate and support information scraps' creation, manipulation and retrieval? We pursued these questions by studying the information scrap practices of 27 knowledge workers at five organizations. Our observations shed light on information scraps' content, form, media, and location. From this data, we elaborate on the typical information scrap lifecycle, and identify common roles that information scraps play: temporary storage, archiving, work-in-progress, reminding, and management of unusual data. These roles suggest a set of unmet design needs in current PIM tools: lightweight entry, unconstrained content, flexible use and adaptability, visibility, and mobility.

Agentic Context Management: Solving Agent Memory and Cost by Treating Them as Lifecycle and Architecture Problems

Production AI agents' failures are less often due to an inability to reason well and more often because they cannot manage what is in their reasoning context: conversation histories, large prompts, large tool definitions, and ballooning tool outputs. Agents drown in their own accumulating history while paying a token cost that grows every turn, producing missing recalls within and across conversations. The incumbent response treats this as a storage-and-retrieval problem. We argue that framing is too narrow. Actively managing what an agent holds in mind is a lifecycle, not merely a store: it spans deciding what to remember, extracting and structuring it, choosing the right store per data type, consolidating and forgetting while preserving provenance, deciding what is relevant now, anticipating what is needed next, and compacting context to a budget without losing what matters. In serious production this operates not over a single user but across an organizational scope hierarchy. We name this discipline Agentic Context Management (ACM) and decompose it into five primitives: architecting, ingesting, scoping, anticipating, and compacting & consolidation. We then make the economic case: naive context accumulation grows token cost quadratically in conversation length, crude summarization buys linear cost at the price of an accuracy cliff, and only validated compaction achieves linear cost with preserved fidelity. We describe a reference implementation, Maximem Synap, that realizes the five primitives as a multi-tenant service and reports 92% on LongMemEval and 93.2% on LoCoMo under the configuration detailed in Section 6. We close with dimensions existing benchmarks do not yet capture, latency, token efficiency, and context-rot resistance, and the frontier of decision-level and organization-level context the category points toward.

CRM Made Simple

Simple yet powerful CRM. Join over 10,000 global businesses and discover an easier way to manage contacts, sales and grow lasting relationships.

User Intents as Living Data — A Descriptive Meta-Framework

A response to Bluesky's User Intents proposal: instead of prescribing a fixed set of intent categories, treat them as living data and use panproto lenses to map between the vocabularies that different communities will inevitably create.

Mem0 - AI Memory Layer for your Agents & Apps | Persistent Context

Mem0 is the memory layer for AI agents and applications, enabling them to learn from past interactions and deliver personalized experiences.