Lore
Knowledge / live index

Blog

34 entries

How we built /share

We launched /share this week. This is how we built it.

EngineeringFeifan ZhouJuly 23, 2026 · 5 min read

How a 3,000-person company got 95% of its employees using AI

A 3,000-person company got 95% of its employees using AI weekly. Their AI Platform Manager explains the three-layer skills system that got them there.

Using AIKaia ColbanJuly 7, 2026 · 3 min read

Lore Changelog #008

Matthew MolinarJuly 23, 20261 min read

Lore's new Workbench enables new forms of AI-assisted sharing and collaboration across threads, skills, and artifacts.

ClaudeMaxxing to 100%: the code, the content, the research, the ops

Kaia ColbanJuly 15, 20266 min read

I'm an ex-consultant with zero engineering background, and I run my whole company on Claude: shipping production code, writing content, doing research, and running ops.

Lore Changelog #007

Matthew MolinarJuly 15, 20261 min read

Lore Studio is a movable macOS dock for working across engineering, knowledge work, and your team's shared thread context.

What shipped in Lore this week

Matthew MolinarJuly 9, 20261 min read

Lore gained decision graphs, clearer sharing, end-to-end private skill sharing, faster search and Ask, and more reliable ingestion.

How I built AI memory (and accidentally built two)

Kaia ColbanJuly 6, 202611 min read

How I gave Claude a persistent memory across Cowork and Claude Code in an afternoon, why it split into two stores, and the bridge file that keeps both in sync in 2026.

Workslop is a context problem, not an AI problem

Kaia ColbanJuly 1, 20265 min read

95% of AI rollouts show no measurable return. The gains are real, but they get eaten by context rot and workslop. Here's why it's a context problem, and the fix.

Lore Updates #0006

Matthew MolinarJune 26, 20261 min read

Publish team skills in one click, collaborate in multiplayer threads, browse full Cowork artifacts, and use a redesigned Lore sidebar.

How Cowork Implements Skills

Feifan ZhouJune 23, 20262 min read

What Claude Desktop is doing under the hood when Cowork discovers, syncs, stores, and injects skills into an agent session.

Stuff nobody tells you about Claude skills

Kaia ColbanJune 22, 20264 min read

Everything I learned about how Claude skills actually work: only the name and description get read, skills live in different places, they fail silently, and a task isn't a skill.

What I learned sharing a Claude skill with my team

Kaia ColbanJune 20, 20263 min read

I built a skill to log my work to Linear, shared it with the team, and found out "everyone has it" is nothing like "everyone uses it."

Lore Changelog #0005

Matthew MolinarJune 19, 20261 min read

Lore gets faster session syncing, CLI skill management, a more stable desktop app, and reliable sign-in across every surface.

Lore vs. Claudebin: sharing a Claude Code session as a URL

Paulina LabaJune 16, 20264 min read

Claudebin pastes a single Claude Code session to a public URL. Lore captures your team's Claude Code, Codex, and Cowork sessions as searchable, permissioned, forkable threads. Here is the difference and when each fits.

Lore Changelog #0004

Matthew MolinarJune 12, 20261 min read

Lore gets a Mac app and a new home at lore.link, alongside persistent thread names, faster browsing, and support for larger sessions.

AI made you faster. It made your team slower.

Kaia ColbanJune 10, 20266 min read

91% of engineers use AI coding tools. Their teammates can see none of the reasoning. Here's why shared AI session context is the real team productivity problem in 2026.

Lore Changelog #0003

Matthew MolinarJune 6, 20261 min read

This week in Lore, conversational threads became native, answers started streaming, workspace sharing opened to everyone, and the app got faster.

Your Team Is in Claude All Day. Where Does That Thinking Go?

Kaia ColbanJune 5, 20265 min read

Your team's best thinking is happening inside Claude Code sessions. Lore captures it, shares it, and makes sure no one builds the same dead end twice.

AI engineering knowledge management: a guide for 2026

Paulina LabaJune 2, 20265 min read

AI coding agents moved your team's reasoning out of docs and into closed terminal sessions. This guide explains what AI engineering knowledge management is, why the old playbook broke, and how to fix it in 2026.

The reasoning behind your code exists. Your team just can't find it.

Kaia ColbanJune 2, 20266 min read

AI coding tools are capturing your team's engineering reasoning in real time. The problem isn't that context is disappearing. It's that nobody can find it.

The best ways to share a Claude Code session (2026)

Paulina LabaJune 2, 20265 min read

A practical, honest rundown of every way to share a Claude Code session in 2026, from screenshots and copy-paste to terminal recordings and Lore, with the trade-offs of each.

How to fork a Claude Code or Codex session

Paulina LabaJune 2, 20264 min read

Forking lets you pick up where someone else's AI coding session left off. Here's how to fork a Claude Code or Codex session with Lore, what a fork actually carries forward, and when to use it.

How to onboard engineers in the AI era

Paulina LabaJune 2, 20264 min read

Onboarding broke when the reasoning behind your codebase moved into AI coding sessions. Here's how to onboard new engineers in 2026 by handing them the sessions, not weeks of Slack archaeology.

Lore vs. Braintrust: what each does and when to use which

Paulina LabaJune 2, 20269 min read

Lore and Braintrust both get filed under "AI observability," but they solve different problems. Lore captures your team's AI coding sessions; Braintrust evaluates the LLM features you ship. Here's how to choose.

Lore vs. LangSmith: what each does and when to use which

Paulina LabaJune 2, 20266 min read

Lore and LangSmith both touch "AI" and "observability," but they watch different things. LangSmith observes the agents inside your product; Lore captures the coding sessions your team runs to build it. Here's how to choose.

Lore vs. ShareGPT: sharing AI chats vs. sharing coding sessions

Paulina LabaJune 2, 20264 min read

ShareGPT shares a ChatGPT conversation as a static snapshot. Lore captures and shares the coding-agent sessions your team runs in Claude Code, Codex, and Cursor. Here's the difference and when each fits.

How to save your favorite prompts from Claude Code and Codex

Paulina LabaJune 2, 20264 min read

Your best prompts are buried in scrollback. Here's how to save and reuse your favorite prompts from Claude Code and Codex with Lore, so the prompt that finally worked is one click away.

Lore Changelog #0002

Feifan ZhouMay 30, 20261 min read

This week in Lore, Ask arrived on web and CLI, sharing gained precise highlights, and skills moved into the Lore plugin.

I Gave Up Solo Building to Join Tanagram.

Kaia ColbanMay 27, 20263 min read

6 months of building alone taught me what I was missing. Here's what happened when I stopped working in the stands and got back on the field.

'AI teams merge 98% more PRs.' What that's doing to your code.

Kaia ColbanMay 26, 20263 min read

AI tools merge PRs 98% faster but introduce quality debt. Learn why traditional velocity metrics fail and 4 better KPIs to measure true engineering productivity.

Lore Changelog #0001

Feifan ZhouMay 23, 20261 min read

This week in Lore, threads gained automatic Regions, on-device upload filters, and direct links to highlighted thread blocks.

Prompt tracking vs. AI session sharing: what to use when

Paulina LabaMay 20, 20264 min read

"Prompt tracking" means two different things in 2026. A practical guide to telling them apart, and choosing the right tool for the one you actually need.

AI made me a faster engineer. It didn't make us a better team.

Kaia ColbanMay 19, 20264 min read

70% of developers say AI helped their personal productivity. Only 17% say it improved team collaboration. That's the most under-discussed gap in 2026 engineering, and it's about to compound.

Managing Tribal Knowledge for Engineers: A Practical Guide

Kaia ColbanMay 15, 20267 min read

Every engineering team has knowledge that lives only in people's heads. Here's why that's gotten worse in the AI era, and what to actually do about it.