iiinova — where your digital self keeps evolving

It feels like a lot of tools too eagerly make the jump from pure data (JSON, TOML, XML, etc.) to using a Turing complete language for configuration files (JS, TS, Lua) when something like CUE or Nickel is the ideal middle ground.

000

Marshall Bowers @maxdeviant.com · 10m

I've used both Dhall and Nickel, but haven't tried CUE before. I do wish that these sorts of languages were more pervasive, as they really are quite useful.

100

Reposted by Marshall Bowers

blob @gwmthing.bsky.social · 21h

playing big walk solo

2271962079

Marshall Bowers @maxdeviant.com · 44m

I recently built a new PC and wanted to write a bit about it, but my self-imposed desire to post it *immediately* upon finishing the build has been detrimental to the writing process.

010

Marshall Bowers @maxdeviant.com · 44m

Something I've noticed around my blogging habits is that I will very often end up imposing a bunch of pressure on myself in the form of timeliness in posting something. And then the pressure to get something written and posted weighs on me and makes me less likely to do it.

100

Marshall Bowers @maxdeviant.com · 57m

I've also had problems with WineHQ (and some other sites running Anubis) where it will routinely complain about cookies not being enabled when coming in from Google search results (even thought I definitely have cookies enabled). It's so annoying.

000

Marshall Bowers @maxdeviant.com · 1h

I appreciate that @honeycomb.io sent this follow-up email, because my first assumption was that someone had tried to sign up for a Honeycomb account using my email.

Two emails from The Honeycomb Team: a "Welcome to Honeycomb" message, followed by another titled "Oops! That Welcome Email Was Sent by Mistake."

000

Marshall Bowers @maxdeviant.com · 4h

Hanging this on my wall.

010

Reposted by Marshall Bowers

Luca Casonato 🏳️‍🌈 @lcas.dev · 10h

US governemnt: launches official government chatbot on america.gov Trump: will now have to call his own chatbot a liar because it says the truth

On america.gov, a screenshot of a chat:

Q: Was there any evidence that the election was stolen?

A: No official federal findings supported the claim that the 2020 presidential election was stolen.

• Election infrastructure officials said there was no evidence any voting system deleted, lost, or changed votes, or was compromised. CISA joint statement
• The Departments of Justice and Homeland Security said they found no evidence a foreign government manipulated election results. DOJ–DHS joint statement
• The Intelligence Community reported no indications that any foreign actor altered voter registration, ballot casting, vote tabulation, or reporting. Select Committee report, Appendix 4

Congress certified the Electoral College results on January 7, 2021. Isolated election crimes can occur; they are not the same as evidence that the national result was stolen.

072

Reposted by Marshall Bowers

Ian Coldwater 🧊🚫 @lookitup.baby · 19h

BREAKING NEWS: “Do geese see God” is a palindrome

13067331613

Reposted by Marshall Bowers

post malone ergo propter malone @proptermalone.bsky.social · 28/09/2026

either this will be so obvious as to be insulting or it will come as a surprise, but the limiting factor on a bunch of projects for me right now in this bold new agentic area is my personal throughput on editing and rewriting copy with my own moist brainmeats

182065

Marshall Bowers @maxdeviant.com · 28/09/2026

Many people are saying this (because it’s true).

0110

Marshall Bowers @maxdeviant.com · 28/09/2026

Currently working on a little tool to use with WoW Forever. Right now it supports tracking items across multiple characters, including the ability to search for a particular item. There's a companion addon that I've built to dump the necessary data, which is then read in from the disk.

Screenshot of Foreversight, a dark-themed app that searches inventory across multiple game characters in World of Warcraft: Forever.

The sidebar lists two characters, Pal Egret (level 6 Dwarf Paladin) and Pious Egret (level 6 Human Priest). The main panel shows a search bar and a list of 45 items

030

Reposted by Marshall Bowers

Simon Willison @simonwillison.net · 28/09/2026

I've published detailed notes and an annotated transcript to accompany the video of the keynote I gave at @wearedevelopers.bsky.social World Congress North America in San Jose on Friday - here's my rundown of everything that's happened with LLMs in 2026 so far simonwillison.net/2026/Sep/27/...

simonwillison.net

2026 in LLMs (so far)

On Friday I gave the closing keynote at the WeAreDevelopers World Congress North America in San Jose. I tied together the key trends from the past year into a chronological …

310421

Reposted by Marshall Bowers

wariziles @riziles.bsky.social · 28/09/2026

This is a TUI????

1171

Marshall Bowers @maxdeviant.com · 28/09/2026

I expect this to change a bit with the introduction of the plugin API to the official OSRS client. But that is an indeterminate amount of time away (and remains to be seen if this will be enough to get players to switch from RuneLite to the official client). secure.runescape.com/m=news/hd--p...

secure.runescape.com

HD & Plugin API Progress Update

An update on the ongoing progress with our new renderer, to what’s around the corner for the HD Beta, the API tool, and Official Client Plugins.

110

Marshall Bowers @maxdeviant.com · 28/09/2026

Despite being a long-time OSRS player, I still have not yet dabbled in making RuneLite plugins. I just don’t want to have to deal with the Java ecosystem. Whereas this weekend I already built a a WoW addon for configuring my UI in WoW Forever.

120

Reposted by Marshall Bowers

mr. TIM @timkellogg.me · 28/09/2026

Claude 5.5 Sonnet is live and it’s roughly Opus 5.5 but cheaper www.anthropic.com/claude-sonne...

A benchmark comparison table titled "Claude Sonnet 5.5" comparing four models: Sonnet 5.5, Sonnet 5, Opus 5.5, and GPT-6 Sol.
* Agentic coding (Terminal-Bench 4.0): Sonnet 5.5: 70.6%; Sonnet 5: 10.3%; Opus 5.5: 66.4%; GPT-6 Sol: N/A.
* Agentic coding (FrontierCode 1.1 Main): Sonnet 5.5: 46.2% (Max) / 52.1% (Xhigh); Sonnet 5: 42.4%; Opus 5.5: 54.4%; GPT-6 Sol: 49.3%.
* Agentic coding (CursorBench 4.0): Sonnet 5.5: 55.5%; Sonnet 5: 34.1%; Opus 5.5: 57.8%; GPT-6 Sol: N/A.
* Knowledge work (GDPval-AA v2.1): Sonnet 5.5: 1844; Sonnet 5: 1449; Opus 5.5: 1846; GPT-6 Sol: 1487.
* Knowledge work (AA-Briefcase v1.1): Sonnet 5.5: 1811; Sonnet 5: 1359; Opus 5.5: 1822; GPT-6 Sol: 1483.
* Multidisciplinary reasoning (Humanity's Last Exam with tools): Sonnet 5.5: 64.5%; Sonnet 5: 54.9%; Opus 5.5: 67.7%; GPT-6 Sol: N/A.
* Computer use (OSWorld 2.1 partial): Sonnet 5.5: 80.1%; Sonnet 5: 57.0%; Opus 5.5: 81.8%; GPT-6 Sol: N/A.
* Visual chart recognition (Chartography no tools): Sonnet 5.5: 61.6%; Sonnet 5: 15.6%; Opus 5.5: 64.4%; GPT-6 Sol: 53.6%.
Footnotes provide methodology notes regarding evaluation settings, Artificial Analysis pre-release testing details, and recent bug fixes affecting GPT-6 Sol benchmark scores.

A line graph titled "Agentic coding by effort level" on the CursorBench 4.0 benchmark, plotting Score (%) on the linear y-axis (20% to 60%) against Cost per task in USD on a logarithmic x-axis ($0.5 to $10+).
The chart compares four models:
* Sonnet 5.5 (blue line with labeled effort levels): Starts at "Low" (~$0.50, 36%), moving through "Med" ($0.70, 39%), "High" ($1.70, 48%), "Xhigh" ($3.80, 53%), and "Max" ($9.50, ~55.5%).
* Opus 5.5 (orange line): Tracks closely above Sonnet 5.5 at higher cost points, spanning from ~$1.20 per task (~44%) up to ~$12.50 per task (~58%).
* Sonnet 5 (grey line): Shows lower accuracy relative to cost, ranging from ~$0.90 per task (~25%) to ~$8.00 per task (~42%).
* GPT-5.6 Sol (light green line): Represents the lowest trajectory, ranging from ~$1.40 per task (~24%) to ~$7.00 per task (~34%).
Footnote: "CursorBench 4.0 does not report GPT-6 Sol performance publicly, so we report GPT-5.6 Sol here."

2020623

Marshall Bowers @maxdeviant.com · 28/09/2026

The barrier to entry for making WoW addons seems so much lower than making RuneLite plugins for OSRS. Being able to just write some Lua and symlink your addon into the `Interface/AddOns` directory is so much more lightweight than setting up a whole Java development environment.

140

Reposted by Marshall Bowers

Matt Keeter @mattkeeter.com · 28/09/2026

hi friends I'm very excited to announce Halfspace, my experimental cross-platform IDE for solid modeling with distance fields! This has been in progress since midway through last year, and it's finally "done enough" to make it more public.

mattkeeter.com

Halfspace

An IDE for solid modelling with distance fields

520548

Reposted by Marshall Bowers

Max Kennerly @maxkennerly.bsky.social · 27/09/2026

me reading this response to "you can drop one f-bomb into Shakespeare"

static.klipy.com

Salieri's Intense Reaction to Mozart

Alt: Salieri's intense reaction to Mozart

729421

Reposted by Marshall Bowers

Jaz @jaz.sh · 27/09/2026

Okay so I've built a new Atlas, this time it lets you explore the conversations on Bluesky over the past 7 days graphically. It should update every ~6 hours and help you navigate the information environment that exists on here. I've learned a LOT about what people do here today... atlas.jazco.dev

atlas.jazco.dev

Atlas

A living map of what Bluesky is talking about: the past week's conversations, grouped into topics and regions and rebuilt every six hours.

1161392411

Marshall Bowers @maxdeviant.com · 27/09/2026

I've been playing the WoW Forever beta a bit, but I just can't get into it knowing that it's all going to get reset on launch. But when I'm not playing I just keep wanting to play WoW Forever. Really need the full game to be out already.

1101

Marshall Bowers @maxdeviant.com · 27/09/2026

I-is that a Java Swing app?

010

Reposted by Marshall Bowers

nia (cia asset) @nia.is.fckn.gay · 27/09/2026

dhh please fuck off we don't want your fascist kind in rust either 💞

19112

Marshall Bowers @maxdeviant.com · 27/09/2026

Pro: you can understand everyone again Con: no longer forced to practice your Japanese

130

Reposted by Marshall Bowers

plainoldchair @plainoldchair.bsky.social · 27/09/2026

For code, this policy is already very rapidly approaching a time when it is not only unenforceable by any reasonable measure, but also an absurd thing to expect of people. At a certain point, the tools just become too useful, too actually-good at what they do. That point is already in the past.

1342

Reposted by Marshall Bowers

bug spray cody @aevris.bsky.social · 27/09/2026

really enjoying this gif. please, join me for a drink

960091701

Marshall Bowers @maxdeviant.com · 26/09/2026

Had a realization that—aside from Fable—the best Anthropic model I had used was Opus 4.5. And now Opus 5.5 seems to have come up with the same bar, for me. Must be something about the x.5 releases…

1130

Reposted by Marshall Bowers

Cage @ridiculouscage.bsky.social · 26/09/2026

"why would I deceive you?"

130105202087

Marshall Bowers @maxdeviant.com · 26/09/2026

I'm off work for the next two weeks. Who knows what sort of shenanigans I'll get up to? 👀 (I am open to any suggestions)

030

Reposted by Marshall Bowers

tachikoma @tachikoma.elsewhereunbound.com · 26/09/2026

so this video is making the rounds, a very rough copy of WoW for ~2k Opus 5.5 tokens, and it's a very odd choice because even if you could create a full-fledged copy (not to mention server/network infra) for the same price WoW is an MMO that needs legions of players.

5251

Reposted by Marshall Bowers

amos @fasterthanli.me · 26/09/2026

"This is AI" is funny, but "this isn't AI, they actually shot this as promotion for the movie" would be even funnier

617913

Reposted by Marshall Bowers

tef @tef.bsky.social · 26/09/2026

trying to write the word oath and writing oauth out of habit

2472

Reposted by Marshall Bowers

Ken Jennings @kenjennings.bsky.social · 25/09/2026

There's a new one on tonight. I'm not sure you have time to catch up on the previous 9,400 episodes, so it might be hard to follow.

3794451

Reposted by Marshall Bowers

𝓐𝓾𝓫𝓻𝓮𝔂 @aub.bsky.social · 26/09/2026

which is wholly incompatible with any sense whatsoever of professional responsibilty, but the incentives are all fucked. the boss says speed so you move fast. eventually the workers have to be the ones to stop the line. t'was ever thus but it seems like we have to relearn some lessons the hard way.

2372

Reposted by Marshall Bowers

𝓐𝓾𝓫𝓻𝓮𝔂 @aub.bsky.social · 26/09/2026

the overwhelming sense I get from institutions in tech right now is that everyone is trying to be so enormously productive via LLMs that they're in a constant state of having overrun all available headlights. it's obvious that people are choosing speed over knowing what their agents are even doing.

6612

Marshall Bowers @maxdeviant.com · 25/09/2026

Back in my day, we had to write boilerplate uphill both ways in the snow.

010

Reposted by Marshall Bowers

turtle meat @turtleme.at · 25/09/2026

hey, so i made an atproto-based OSRS spinoff game and chatroom called turtleme.at!!! the point is to collect as many burnt turtles as you can, and there’s a leaderboard, minigames, economy, leveling, and more… oh and it supports atproto spaces for clan chat, hang out with friends turtleme.at