iiinova — where your digital self keeps evolving

ParallelPilot: Supporting Coordination and Monitoring in Parallel AI Coding arxiv.org/abs/2609.33113

Looks like Gemini Argon takes more tokens to be able to deliver the same level of intelligence as Opus 5.5 (high; not max) so it would be slower and more expensive than Opus if not for the temporary 50% discount artificialanalysis.ai/models/gemin...

000

Shawn Simister @narphorium.com · 25/09/2026

"I would have written a shorter report, but I didn't have enough tokens"

1362

Shawn Simister @narphorium.com · 20/09/2026

Not yet. In that case Prolog would be the system one model. What I've been trying to get to is a combination of the search and backtracking algorithm from prolog with the natural language understanding of LLMs but until now its just been too slow

011

Shawn Simister @narphorium.com · 19/09/2026

But I should also just clean up the code and open source it so you can try it out for yourself

160

Shawn Simister @narphorium.com · 19/09/2026

Yeah, I decided just to record a quick video. Didn't realize this many people would reshare it. If you look closely, your screenshot shows "step 16 of 84". I can assure you that by step 34 it finds all the grandpa relations

230

Shawn Simister @narphorium.com · 18/09/2026

That would be awesome but this version just does simple exhaustive search through the facts. I have some ideas for smarter search. I think it would probably need a proper generative model to predict the best search queries though.

040

Shawn Simister @narphorium.com · 18/09/2026

Jev decides if two statements can be unified. So it can do fuzzy matching on natural language statements

290

Shawn Simister @narphorium.com · 18/09/2026

Jev doesn't generate text, so I used Claude Haiku to turn the question into a logic query first. After that, this query took 100 Jev calls in 8.5 seconds, for roughly a quarter of a cent.

1200

Shawn Simister @narphorium.com · 18/09/2026

To match a goal against a fact or a rule, the interpreter asks Jev questions like: Could "Abe is the father of Y" and "Abe is Homer's father" state the same fact? Jev answers with a probability. The rest is ordinary proof search, written in code.

1180

Shawn Simister @narphorium.com · 18/09/2026

Jev can't reason. So I built an interpreter that reasons for it. It runs a logic language where the facts and rules are plain English.

71149

Shawn Simister @narphorium.com · 17/09/2026

Thanks! I'll have to test that out and see how they compare

120

Reposted by Shawn Simister

Tudor Girba @tudorgirba.com · 17/09/2026

"Software engineering has changed radically. It is now mainly about reading code. But reading code does not scale. We need to delegate to AI." I am glad you now recognize code reading as a problem. However, the broader problem statement is not quite correct and neither is the conclusion.

software engineering has not changed radically.

it has always been mainly about understanding code.

but reading code does not scale.

we need to augment our understanding through contextual tools.

183

Shawn Simister @narphorium.com · 15/09/2026

I think these walkthroughs need to be part of a larger interactive loop and not just something you review at the end. People want to skim over some concepts and dig deep into the details in other areas. One of the challenges right now is figuring out the right levels of semantic zoom.

010

Shawn Simister @narphorium.com · 14/09/2026

We can use AI to build tools for exploration and understanding too bsky.app/profile/narp...

100

Shawn Simister @narphorium.com · 14/09/2026

Love my insanely large monitor but the resolution (5120 x 1440) isn’t really that much greater than other sanely large monitors

110

Shawn Simister @narphorium.com · 14/09/2026

TIL if you export CLAUDE_CODE_OAUTH_TOKEN it also disables Chrome MCP so Claude doesn't know how to fix simple CSS bugs anymore. Fun times code.claude.com/docs/en/chrome

code.claude.com

Use Claude Code with Chrome - Claude Code Docs

Connect Claude Code to your Chrome browser to test web apps, debug with console logs, automate form filling, and extract data from web pages.

110

Shawn Simister @narphorium.com · 10/09/2026

This is great. I’ve been circling a similar idea bsky.app/profile/narp...

020

Shawn Simister @narphorium.com · 08/09/2026

I've been doing that a lot too. Have you tried Chopin over Tailscale?

100

Reposted by Shawn Simister

Andy Matuschak @andymatuschak.org · 08/09/2026

Been tinkering with annotation "fusion": annotate a paper with ink; take a photo; markup is segmented, turned into structured data, and re-super-imposed on the clean original PDF.

2371

Shawn Simister @narphorium.com · 08/09/2026

This is inspiring. I also experimented with the idea of hand-drawn notes being a compressed idea which the tool unpacks into somethingore detailed which can be edited and even executed github.com/narphorium/s...

github.com

GitHub - narphorium/sketch2canvas: Sketch2Canvas used Claude to automatically turn sketches into executable Obsidian canvases

Sketch2Canvas used Claude to automatically turn sketches into executable Obsidian canvases - narphorium/sketch2canvas

021

Reposted by Shawn Simister

Observable @observablehq.com · 01/09/2026

We reimagined how computational notebooks could work with agents as the primary interface. Our new agent-first notebooks, called chats, are radically different from our human-first notebooks and yet seamlessly interoperate with them.

A screenshot of an Observable “chat” interface. Several interface elements are labeled, notably: a prompt, upload file button, and model chooser at the bottom.

161

Shawn Simister @narphorium.com · 01/09/2026

Chances feel slim but I agree bsky.app/profile/narp...

010

Shawn Simister @narphorium.com · 31/08/2026

Nice writeup here about how coding agents can score similarly on benchmarks but operate very differently behind the scenes blog.jetbrains.com/junie/2026/0...

450

Shawn Simister @narphorium.com · 29/08/2026

I built my first PC before I had smartphone so when I took it apart for repairs I had no computer and no internet. It was a long time ago but I still remember that feeling

010

Shawn Simister @narphorium.com · 28/08/2026

Yes! I’ve been working in this too bsky.app/profile/narp...

030

Reposted by Shawn Simister

Jeremy Lewi @jeremy.lewi.us · 20/08/2026

With Codex/Claude the bottleneck in software development is now expressing your intent. Writing the code is no longer the rate limiting step. This completely upends software development. Take issue trackers. The key problem they were solving was allocation of scarce coding resources to features.

131

Shawn Simister @narphorium.com · 17/08/2026

xanadu or xanadu not. there is no try

030

Reposted by Shawn Simister

Stuart Gray @sgray.bsky.social · 12/08/2026

An interesting paper with implications for skill extraction, continual learning, and token efficiency - it provides even more incentive & value from running user-specific task evals because you can extract skills from them. arxiv.org/abs/2608.07885

arxiv.org

Reason Wide, Not Deep: Amortizing the Reasoning Premium into Distilled Skills

Reasoning modes of language models outperform their non-reasoning counterparts on multi-step agentic tasks, but pay a 3-6x premium in output tokens on every episode -- much of it spent re-deriving pro...

151

Reposted by Shawn Simister

Tom Silver @tomssilver.bsky.social · 09/08/2026

This week's #PaperILike is "People construct simplified mental representations to plan" (Ho et al., Nature 2022). People generate ("construe") simplified task representations on-the-fly. (Robots should too!) Also: awesome first paragraph. PDF: arxiv.org/abs/2105.06948

arxiv.org

People construct simplified mental representations to plan

One of the most striking features of human cognition is the capacity to plan. Two aspects of human planning stand out: its efficiency and flexibility. Efficiency is especially impressive because plans...

6418

Shawn Simister @narphorium.com · 05/08/2026

paper maps -> MapQuest -> Maps on your phone + Streetview

050

Shawn Simister @narphorium.com · 02/08/2026

really enjoying this setup for syntopic reading with Fable

001

Shawn Simister @narphorium.com · 30/07/2026

Keyset on the left, mouse on the right, like Engelbart's setup in the Mother of All Demos

020

Shawn Simister @narphorium.com · 30/07/2026

My transparent Creator Micro keys arrived, so now I can make atelier.dev decision points map to physical buttons

160

Shawn Simister @narphorium.com · 23/07/2026

TodoWrite was replaced by TaskCreate earlier this year

010

Shawn Simister @narphorium.com · 18/07/2026

"the phoenix must be reborn with its empty spaces intact — intent includes the intent to remain open, and “immortality of intent” is only half-achieved by a system that can’t carry a dodge through the fire." wtf you talking about fable? 😂

001

Shawn Simister @narphorium.com · 15/07/2026

I guess what I really want is a portable Dynalamp dynamicland.org/archived-med...

000

Shawn Simister @narphorium.com · 14/07/2026

Just realized that my portable table lamp would make the perfect vessel for a voice-powered AI assistant

130

Shawn Simister @narphorium.com · 14/07/2026

🤔 very tempting jsaux.com/pages/steam-...

110

Shawn Simister @narphorium.com · 26/06/2026

Thanks Wil! Great to connect with more FoC people on here

010