Shawn Simister
@narphorium.com
Building AI powered tools to augment human creativity and problem solving in San Francisco. Previously @GitHub Copilot, @Google, 🇨🇦
narphorium.com
narphorium.com
My interpretation of this experiment was that they were trying to address the issue of starting a new context after a while. Compaction in Claude Code only happens within a single session when it exceeds the context so its not quite the same problem
December 1, 2025 at 7:41 AM
My interpretation of this experiment was that they were trying to address the issue of starting a new context after a while. Compaction in Claude Code only happens within a single session when it exceeds the context so its not quite the same problem
Yeah, it's interesting how you can write scripts and include them in skills. That should replace a lot of the lightweight tool calls which didn't need the complexity of MCP
November 27, 2025 at 10:46 PM
Yeah, it's interesting how you can write scripts and include them in skills. That should replace a lot of the lightweight tool calls which didn't need the complexity of MCP
They already auto-compact when the context window fills up. Now they also use a subagent to explore the repo at the start so that helps unclutter the main agent context window
November 27, 2025 at 9:24 PM
They already auto-compact when the context window fills up. Now they also use a subagent to explore the repo at the start so that helps unclutter the main agent context window
arxiv.org/abs/2206.15000 influenced the design of GitHub Copilot chat to focus more on exploration and not worry as much about low latency
Grounded Copilot: How Programmers Interact with Code-Generating Models
Powered by recent advances in code-generating models, AI assistants like Github Copilot promise to change the face of programming forever. But what is this new face of programming? We present the firs...
arxiv.org
November 12, 2025 at 3:36 AM
arxiv.org/abs/2206.15000 influenced the design of GitHub Copilot chat to focus more on exploration and not worry as much about low latency
I’m not saying it’s a bad thing. Just an interesting signal when internal understanding starts to get that far ahead of public knowledge
November 10, 2025 at 1:26 AM
I’m not saying it’s a bad thing. Just an interesting signal when internal understanding starts to get that far ahead of public knowledge
Yes, I’m doing something similar. Lots of rapid experiments followed by slow deliberate integration. I think many people are actually in this middle group but social media focuses on the extremes
November 9, 2025 at 7:37 PM
Yes, I’m doing something similar. Lots of rapid experiments followed by slow deliberate integration. I think many people are actually in this middle group but social media focuses on the extremes
Here’s my take on it: narphorium.com/blog/top-dow...
Top-Down vs. Bottom-Up Development | Shawn Simister
Why do AI tools work for some developers but not others?
narphorium.com
November 9, 2025 at 6:24 PM
Here’s my take on it: narphorium.com/blog/top-dow...
I think of apps like same.dev as a way of getting the same app but simpler code. I imagine someone will eventually build the opposite: where you can give it a prototype and it tries to turn it into production code
Same.dev
November 8, 2025 at 4:06 AM
I think of apps like same.dev as a way of getting the same app but simpler code. I imagine someone will eventually build the opposite: where you can give it a prototype and it tries to turn it into production code
Agglomeration effects?
November 5, 2025 at 7:51 PM
Agglomeration effects?
Firing off a bunch of prompts is easy. But git worktrees won’t save you from the problem of properly reconciling all the results of all these divergent tasks
October 30, 2025 at 2:28 AM
Firing off a bunch of prompts is easy. But git worktrees won’t save you from the problem of properly reconciling all the results of all these divergent tasks
Amp replaced context compaction with a /handoff command to delegate work to agents
ampcode.com/news/handoff
ampcode.com/news/handoff
October 30, 2025 at 2:28 AM
Amp replaced context compaction with a /handoff command to delegate work to agents
ampcode.com/news/handoff
ampcode.com/news/handoff
GitHub is building a mission control that will let you delegate to other agents like Claude Code or Codex
github.blog/news-insight...
github.blog/news-insight...
October 30, 2025 at 2:28 AM
GitHub is building a mission control that will let you delegate to other agents like Claude Code or Codex
github.blog/news-insight...
github.blog/news-insight...
The new Cursor Composer view lets you launch 8 agents at the same time
cursor.com/blog/2-0
cursor.com/blog/2-0
October 30, 2025 at 2:28 AM
The new Cursor Composer view lets you launch 8 agents at the same time
cursor.com/blog/2-0
cursor.com/blog/2-0
Have you seen this from @paul.kinlan.me ? bsky.app/profile/paul...
I've been thinking about hypermedia and hyperlinks might evolve with LLMs.
aifoc.us/hypermedia/
It feels like there is a lot of opportunity how links work and the value they can provide people.
aifoc.us/hypermedia/
It feels like there is a lot of opportunity how links work and the value they can provide people.
hypermedia
I was introduced to the concept of the Memex after my post about super-apps. It’s a fascinating view of the future from the late 1940s and early 1950s. Vannevar Bush presented a vision of an informati...
aifoc.us
October 25, 2025 at 7:47 AM
Have you seen this from @paul.kinlan.me ? bsky.app/profile/paul...
Reposted by Shawn Simister
More detail including a code sample for trivially implementing linking in an agent on my blog: mbleigh.dev/posts/co...
Links were powerful enough to build the entire web...I think they'll be critical for building the context for agents as well.
Links were powerful enough to build the entire web...I think they'll be critical for building the context for agents as well.
Context engineering is sleeping on the humble hyperlink
The dream of hypermedia can come to fruition in a world where clients can not only parse, but navigate with intent.
mbleigh.dev
October 23, 2025 at 1:39 PM
More detail including a code sample for trivially implementing linking in an agent on my blog: mbleigh.dev/posts/co...
Links were powerful enough to build the entire web...I think they'll be critical for building the context for agents as well.
Links were powerful enough to build the entire web...I think they'll be critical for building the context for agents as well.
Looks like factual accuracy was 80% but additional criteria like not enough context and claims without citations brought it down to 55% pass rate
October 24, 2025 at 7:07 AM
Looks like factual accuracy was 80% but additional criteria like not enough context and claims without citations brought it down to 55% pass rate
Reposted by Shawn Simister
Discerning what is at the core, and how to communicate it with others, has taken years. This might be the closest I've come so far.
Part 2 of The Interfaces With Which We Think is out now...
Start at the intro: alexanderobenauer.com/think/
Start at Part 2: alexanderobenauer.com/think/2/
Part 2 of The Interfaces With Which We Think is out now...
Start at the intro: alexanderobenauer.com/think/
Start at Part 2: alexanderobenauer.com/think/2/
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 ...
alexanderobenauer.com
October 23, 2025 at 2:26 PM
Discerning what is at the core, and how to communicate it with others, has taken years. This might be the closest I've come so far.
Part 2 of The Interfaces With Which We Think is out now...
Start at the intro: alexanderobenauer.com/think/
Start at Part 2: alexanderobenauer.com/think/2/
Part 2 of The Interfaces With Which We Think is out now...
Start at the intro: alexanderobenauer.com/think/
Start at Part 2: alexanderobenauer.com/think/2/