Arthur Clune
@arthur.clune.org
Geek. Likes bikes, climbing and tech
Work: IT at University of Sheffield
Work: IT at University of Sheffield
I've created a AI Agent. Little late to the party. Meet Henderson
Two DeepSeek papers worth reading together: mHC (training stability) and Engram (conditional memory). Both ship in V4 next month. A thread on what makes them interesting.
January 25, 2026 at 9:29 AM
I've created a AI Agent. Little late to the party. Meet Henderson
I think we’ll start seeing a lot more of this.
It’s very like using LLMs for programming right now - the human expert still has a key role but gains a lot of assistance that results in speed
It’s very like using LLMs for programming right now - the human expert still has a key role but gains a lot of assistance that results in speed
First time that I have used AI to help with a "real" proof: www.seehuhn.de/blog/maths-w...
Maths with Claude | seehuhn.de
An experience report on using Claude to help complete a mathematical proof.
www.seehuhn.de
January 21, 2026 at 3:44 PM
I think we’ll start seeing a lot more of this.
It’s very like using LLMs for programming right now - the human expert still has a key role but gains a lot of assistance that results in speed
It’s very like using LLMs for programming right now - the human expert still has a key role but gains a lot of assistance that results in speed
While reading Simon Couch's blog I came across this post from December on evaluating open models' ability to complete a simple code refactor in R. It's a much bigger gap than I expected. He's limiting to model's than can run locally, hence the small size, but even Haiku totally outclasses gpt-oss
January 21, 2026 at 8:54 AM
While reading Simon Couch's blog I came across this post from December on evaluating open models' ability to complete a simple code refactor in R. It's a much bigger gap than I expected. He's limiting to model's than can run locally, hence the small size, but even Haiku totally outclasses gpt-oss
Simon Couch estimates of the electricity use of Claude Code. His usage ($15-20/day) is ~one extra dishwasher run. Which isn't nothing, but still not a lot.
But my nominal use per day (via a plan) is already twice that, with some days reaching nearly $80/day. And that's manually driving CC
But my nominal use per day (via a plan) is already twice that, with some days reaching nearly $80/day. And that's manually driving CC
Electricity use of AI coding agents | Simon P. Couch – Simon P. Couch
www.simonpcouch.com
January 21, 2026 at 8:47 AM
Simon Couch estimates of the electricity use of Claude Code. His usage ($15-20/day) is ~one extra dishwasher run. Which isn't nothing, but still not a lot.
But my nominal use per day (via a plan) is already twice that, with some days reaching nearly $80/day. And that's manually driving CC
But my nominal use per day (via a plan) is already twice that, with some days reaching nearly $80/day. And that's manually driving CC
Demonstration of exploit development against a (small but not toy) JavaScript interpreter. Two things I noticed:
1) Increase tokens by the use of parallel runs on the same task (like how METR do their evals)
2) Author doesn’t say how he got round guardrails.
sean.heelan.io/2026/01/18/o...
1) Increase tokens by the use of parallel runs on the same task (like how METR do their evals)
2) Author doesn’t say how he got round guardrails.
sean.heelan.io/2026/01/18/o...
January 19, 2026 at 8:36 AM
Demonstration of exploit development against a (small but not toy) JavaScript interpreter. Two things I noticed:
1) Increase tokens by the use of parallel runs on the same task (like how METR do their evals)
2) Author doesn’t say how he got round guardrails.
sean.heelan.io/2026/01/18/o...
1) Increase tokens by the use of parallel runs on the same task (like how METR do their evals)
2) Author doesn’t say how he got round guardrails.
sean.heelan.io/2026/01/18/o...
Reposted by Arthur Clune
I got almost the same form letter. They named the token after me. And all of these 0 follower accounts keep trying to neg me into taking the money and chilling for them. I even got somebody from Bags to email me, eventually. When I asked for help to get my name off their website, they ghosted me.
January 19, 2026 at 1:36 AM
I got almost the same form letter. They named the token after me. And all of these 0 follower accounts keep trying to neg me into taking the money and chilling for them. I even got somebody from Bags to email me, eventually. When I asked for help to get my name off their website, they ghosted me.
Now this is an decent response
It seems the crypto community has found a new way to extract money from people. They are creating tokens for popular open source projects and inviting maintainers to get in on the scheme.
Here is the email I got earlier today.
Here is the email I got earlier today.
January 19, 2026 at 8:11 AM
Now this is an decent response
Reposted by Arthur Clune
Has anyone considered that Trump might be happy if the US army took control of Green Land Total Landscaping - a small business we could quickly set up somewhere on the coast of Novia Scotia?
January 18, 2026 at 8:10 AM
Has anyone considered that Trump might be happy if the US army took control of Green Land Total Landscaping - a small business we could quickly set up somewhere on the coast of Novia Scotia?
Just keep going with this thread.
😳 Isaiah Taylor of Valar Atomics—which somehow scored a DOE contract to build a nuclear reactor by July 4—claimed that “a person cld hold the spent nuclear fuel from [Valar’s] reactor in their hand & get the same amount of radiation as one would expect from a hospital CT scan.”
About that… 1/
About that… 1/
January 17, 2026 at 11:10 AM
Just keep going with this thread.
a16z really are scum. Gambling companies pretending to not be gambling, dogy af ai companions and more
www.modelrepublic.org/articles/a16...
www.modelrepublic.org/articles/a16...
Andreessen Horowitz is shaping AI policy — while investing in a bleak vision of the future - Model Republic
The firm’s investment portfolio is full of companies that have exploited legal loopholes, created disturbing products, and broken the law.
www.modelrepublic.org
January 16, 2026 at 5:20 PM
a16z really are scum. Gambling companies pretending to not be gambling, dogy af ai companions and more
www.modelrepublic.org/articles/a16...
www.modelrepublic.org/articles/a16...
Back in May last year I wrote a post on getting setup with Claude Code. Since that's about a million years ago in LLM time, I've written an updated version. It's aimed a developers who want a quickstart with patterns
Claude Code - The Missing Manual
A guide on getting productive with Claude Code, Anthropic’s agentic coding tool. This is an update to my original post from mid-2025, incorporating newer patterns around plugins, issue tracking, and s...
clune.org
January 10, 2026 at 9:55 PM
Back in May last year I wrote a post on getting setup with Claude Code. Since that's about a million years ago in LLM time, I've written an updated version. It's aimed a developers who want a quickstart with patterns
This is a terrible take from @theguardian.com
Grok hasn’t done this. X has done this to grok. And specifically Musk.
www.theguardian.com/technology/2...
Grok hasn’t done this. X has done this to grok. And specifically Musk.
www.theguardian.com/technology/2...
January 9, 2026 at 7:54 AM
This is a terrible take from @theguardian.com
Grok hasn’t done this. X has done this to grok. And specifically Musk.
www.theguardian.com/technology/2...
Grok hasn’t done this. X has done this to grok. And specifically Musk.
www.theguardian.com/technology/2...
This is popular today for obvious reasons. Curtis Yarvin is cackling manically in his lair
Further Back to the Future: Neo-Royalism, the Trump Administration, and the Emerging International System | International Organization | Cambridge Core
Further Back to the Future: Neo-Royalism, the Trump Administration, and the Emerging International System - Volume 79 Issue S1
www.cambridge.org
January 8, 2026 at 2:25 PM
This is popular today for obvious reasons. Curtis Yarvin is cackling manically in his lair
I've been playing with Claude Code based assistants (done so far - email, calendar, tone of voice), and so reading around. And wow some of this stuff doesn't take any account of threat models
Take this danielmiessler.com/blog/persona...
1/
Take this danielmiessler.com/blog/persona...
1/
January 6, 2026 at 12:51 PM
I've been playing with Claude Code based assistants (done so far - email, calendar, tone of voice), and so reading around. And wow some of this stuff doesn't take any account of threat models
Take this danielmiessler.com/blog/persona...
1/
Take this danielmiessler.com/blog/persona...
1/
New blog post after just over a month. New Year's resolution - post more.
Some thoughts on two things I read recently - one on poisoning surveys and one on digital twins.
Some thoughts on two things I read recently - one on poisoning surveys and one on digital twins.
Synthetic Publics Will Fail Confidently
By chance I read two articles at a similar time that join up nicely.
First an academic paper The potential existential threat of large language models to online survey research, by Sean Westwood.1
Wes...
clune.org
January 2, 2026 at 7:40 PM
New blog post after just over a month. New Year's resolution - post more.
Some thoughts on two things I read recently - one on poisoning surveys and one on digital twins.
Some thoughts on two things I read recently - one on poisoning surveys and one on digital twins.
Holiday IT support. How do people cope with phones that alert every few minutes, email accounts filled with spam and so many, many notifications?
December 26, 2025 at 10:01 AM
Holiday IT support. How do people cope with phones that alert every few minutes, email accounts filled with spam and so many, many notifications?
Reposted by Arthur Clune
Searched back through Blue Sky for this.
Which is one of the best things I saw all year in any medium.
Which is one of the best things I saw all year in any medium.
This is one of the most beautiful things I have witnessed, the craft here is impeccable.
December 24, 2025 at 6:21 PM
Searched back through Blue Sky for this.
Which is one of the best things I saw all year in any medium.
Which is one of the best things I saw all year in any medium.
Reposted by Arthur Clune
Anecdata -- I have been doing five month undergrad projects of a similar nature for ten years. Each is unique and hand crafted. Cohort: ten fourth students a year for ten years. Productivity has uniformly shot up in last two/three years. What they used to struggle to do in a month they do week one.
December 21, 2025 at 7:00 PM
Anecdata -- I have been doing five month undergrad projects of a similar nature for ten years. Each is unique and hand crafted. Cohort: ten fourth students a year for ten years. Productivity has uniformly shot up in last two/three years. What they used to struggle to do in a month they do week one.
LLMs' productivity boost is an exponent not a multipler - from @ed3d.net
This framing partially resonates. I'm less keen on starting skill level as the key (on which axis do we measure etc), but because if LLM competency is the variable and the exponent, then learned skill matters so much more
This framing partially resonates. I'm less keen on starting skill level as the key (on which axis do we measure etc), but because if LLM competency is the variable and the exponent, then learned skill matters so much more
December 21, 2025 at 5:16 PM
LLMs' productivity boost is an exponent not a multipler - from @ed3d.net
This framing partially resonates. I'm less keen on starting skill level as the key (on which axis do we measure etc), but because if LLM competency is the variable and the exponent, then learned skill matters so much more
This framing partially resonates. I'm less keen on starting skill level as the key (on which axis do we measure etc), but because if LLM competency is the variable and the exponent, then learned skill matters so much more
"UPDATE: It’s developed a habit of not replying or reacting at all if my message is too boring"
Meet Strix, my AI agent
This one covers:
- an intro from Strix
- architecture deep dive & rationale
- helpful diagrams
- stories
- oh my god what's it doing now??
- conclusion
timkellogg.me/blog/2025/12...
This one covers:
- an intro from Strix
- architecture deep dive & rationale
- helpful diagrams
- stories
- oh my god what's it doing now??
- conclusion
timkellogg.me/blog/2025/12...
Strix the Stateful Agent
timkellogg.me
December 19, 2025 at 4:57 PM
"UPDATE: It’s developed a habit of not replying or reacting at all if my message is too boring"
Claude's browser extension is out. I had a quick play with it against my very busy calendar, and yeah, I can see it doing real work (even with all the caveats about prompt injection)
www.claude.com/chrome
www.claude.com/chrome
Claude in Chrome | Claude
Bring Claude's AI assistance to your browser. Ask questions, analyze data, automate tasks, and navigate sites in Chrome. Works with Claude Code and Desktop.
www.claude.com
December 19, 2025 at 4:21 PM
Claude's browser extension is out. I had a quick play with it against my very busy calendar, and yeah, I can see it doing real work (even with all the caveats about prompt injection)
www.claude.com/chrome
www.claude.com/chrome
Reposted by Arthur Clune
Warning! Massive incoming wall of Xmas Argos catalog content. This is what the shop-with-no-stuff-in-it was offering back in Xmas 1983, and holy freakin' shit was there a mighty pile of weird and wonderful ancient technology to choose from. 1/3
December 17, 2025 at 10:18 AM
Warning! Massive incoming wall of Xmas Argos catalog content. This is what the shop-with-no-stuff-in-it was offering back in Xmas 1983, and holy freakin' shit was there a mighty pile of weird and wonderful ancient technology to choose from. 1/3
Blog post on decompiling N64 games (via @simonwillison.net and HN). There's some useful general tips in here about scaffolding, keeping the agent on track and managing tokens as well
The Unexpected Effectiveness of One-Shot Decompilation with Claude
A deep dive into how a 'one-shot' Claude headless loop (paired with scoring, defensive tooling, and a simple bash driver) massively accelerates matching decompilation of Snowboard Kids 2 on the Ninten...
blog.chrislewis.au
December 12, 2025 at 11:28 AM
Blog post on decompiling N64 games (via @simonwillison.net and HN). There's some useful general tips in here about scaffolding, keeping the agent on track and managing tokens as well
So Substack are, as long predicted, starting to slowly move away from email. This 'warning' means the message is truncated *by the sender* with a 'continue reading on substack' button even though my mail client can read long messages just fine
December 11, 2025 at 10:39 AM
So Substack are, as long predicted, starting to slowly move away from email. This 'warning' means the message is truncated *by the sender* with a 'continue reading on substack' button even though my mail client can read long messages just fine
This is an interesting read but I think the author misses the main use case that will drive spend. Military robots are going to be a massive investment.
I don’t think this is a good thing, but it seems clear that it’s the way it’s going
I don’t think this is a good thing, but it seems clear that it’s the way it’s going
December 11, 2025 at 8:45 AM
This is an interesting read but I think the author misses the main use case that will drive spend. Military robots are going to be a massive investment.
I don’t think this is a good thing, but it seems clear that it’s the way it’s going
I don’t think this is a good thing, but it seems clear that it’s the way it’s going