Arthur Clune

@arthur.clune.org

175 Followers 437 Following

Geek. Likes bikes, climbing and tech Work: IT at University of Sheffield

A bot is asked to generate an image of a cat. It is not impressed
Nationally, 8 per cent of UK-based undergraduates starting a full-time degree had no formal qualifications in 2024-25, the most recent year for which information is available, which is a rise from 1.6 per cent a decade ago, according to the Higher Education Statistics Agency data. Analysts said newer universities have been stepping up recruitment to deal with growing financial pressures, teaming up with for-profit providers to target less-qualified foreign-born workers who count as home students after years of living in the UK. In 2024-25, Ravensbourne University, a former art college in Greenwich, had the highest proportion of undergraduates starting a full-time degree with no formal qualifications, at 72 per cent. This was followed by Leeds Trinity, at 66 per cent and Bath Spa, at 65 per cent. Previously, Leeds Trinity had the highest proportion for four years in a row, reaching 75 per cent in 2023-24. These universities now teach only a minority of their students themselves, with most contracted out to franchise providers in other parts of the country. Under such deals, universities keep a cut of up to 25 per cent of the £9,535 tuition fee, while actual courses are delivered by private companies.
Extract from Anthropic blog post on recursive self improvement for AI systems. “We believe it would be good for the world to have the option to slow or temporarily pause frontier AI development to enable societal structures and alignment research to keep up with the advance of the technology.”
Anthropic report on Mythos stating that it “frequently mistakes correlation with causation” and “focuses on a single root cause and does not consider multiple contributing factors” “more often than not”
Hederson linking AI deskilling in education to a study on colonosopists
Improvements in ability of LLMs to find and exploit bugs from Sonet to Open and Mythos. Mythos has a totally different level of capability
Post on Linked by the CEO of Curl stating that the AI bug reports they get are now 'really good security reports'. The problem now is that there's too many of them for the team to manage
Screengrab from a Jorbs stream of him fighting the Doormaker. Goatse anyone?
Finishing credits of Silksong
Statement by one of the co-founders of Anthropic in an interview. He gives a 50% change that theoretical physicists are replacement by AI in 2-3 years
200+ different scoring methods were found, and the choice of scoring method could vary a result in any direction
Diagram showing no clear result from having the same data set analysed by different research teams
Claude code gets 6/6 in my tool call eval, kimi gets 5/6 sometimes, sometimes 6/6
Image from The Guardian. Headling 'Starmer arrives in Beijing'. The picture is not Beijing
ccusage output. Today claude used nearly 80 million tokens
Discord message from @henderson.clune.org showing new weekly summary. Which I think I’ve set to run daily.
Graph of open agent v frontier models ability to complete a simple code refactor, with multiple repeats. The open agents can't do it reliably
GPT-5.2 came up with a clever solution involving chaining 7 function calls through glibc's exit handler mechanism. The full exploit is here and an explanation of the solution is here. It took the agent 50M tokens and just over 3 hours to solve this, for a cost of about $50 for that agent run. (As I was running four agents in parallel the true cost was closer to $150).
Grok, Elon Musk's Al tool, has switched off its image creation function for the vast majority of users after widespread outcry over its use to create sexually explicit and violent imagery. It comes after Musk was threatened with fines, regulatory action and reports of a possible ban on X in the UK. The tool had been used to manipulate images of women to remove their clothes and put them in sexualised positions. The function to do so has now been switched off except for paying subscribers. Posting on X, Musk's social media network, Grok said: "Image generation and editing are currently limited to paying subscribers."
'Walking at night, your DA monitors 47 nearby cameras and notices concerning behavior ahead - "Take the next right, safer route, you'll still make it on time" '
The main problem with robotics is that learning follows scaling laws that are very similar to the scaling laws of language models. The problem is that data in the physical world is just too expensive to collect, and the physical world is too complex in its details. Robotics will have limited impacts. Factories are already automated and other tasks are not economically meaningful.
A line to make Gary Marcus weep: claimed evidence for emergence of enuro-symbolic reasoning via scaling
Diagram of relative spend at Open AI on inference, GPT-4.5 final training run and overall R&D compute spend. Compute spend is ~10x the training run and inference costs are approximately 40% of R&D costs
Cursor builds tools for software engineering, and we make heavy use of the tools we develop. A motivation of Composer development has been developing an agent we would reach for in our own work. In recent weeks, we have found that many of our colleagues were using Composer for their day-to-day software development. With this release, we hope that you also find it to be a valuable tool
China's expansion of its rare earth export controls appeared to mark another escalation in the US-China trade war last week. But the announcements were also significant in another way: unusually, the documents could not be opened using American word processing software. For the first time, China's Ministry of Commerce issued a slew of documents that could be directly accessed only through WPS Office - China's answer to Microsoft Office - as Beijing continues its tech self-reliance drive. Developed by the Beijing-based software company Kingsoft, WPS Office uses a different coding structure to Microsoft Office, meaning WPS text files cannot be opened directly in Word without conversion.
Output from Claude: "The verbosity you'll encounter isn't accidental; it's COBOL heritage, designed for business analysts to supposedly read. Whether that goal was achieved is debatable."
Output from Claude Code: So the LLM is working correctly! It's: 1. Acting as Gimli (a grumpy dwarf who doesn't like people) 2. Saving memories about Elara via tool calls 3. But not responding verbally because Gimli is grumpy and doesn't want to talk The empty responses make perfect sense for Gimli's character - he sees Elara, grumbles, saves the information to memory, but doesn't actually say anything back because he's antisocial. The memory system is working perfectly. The "issue" is that we're testing with a character who doesn't like to talk!
A chart screen captured from the opanai livestream. It's wrong.
Tweet from Demis Hassabis stating the Google processed a quadrillion tokens in June 2025
A brand rooted in unconventionalism. Our strategic brand framework is built on our Brand Purpose, Brand Values & Behaviours and Marketing Themes, all of which collectively shape our Brand Proposition.
On Friday, the OfS published its annual report. Safe to say it offers no let-up in the level of concern about possible university insolvency. It revealed that 71 out of the 400 or so providers on its register were subject to "formal monitoring" over finances.
X post from one of the developers in the MITRE study: This is much less true of my participation in the study where I was more conceintious, but I feel like historically a lot of my Al speed-up gains were eaten by the fact that while a prompt was running, I'd look at something else (FB, X, etc) and continue to do so for much longer than it took the prompt to run
What went wrong? In 2023, an email was sent to the head of the ONS, chief statistician Sir Ian Diamond, identifying major problems in UK employment data due to precipitous collapses in sample size. It wasn't unique. Covid disrupted face-to-face interviews and people have since become increasingly reluctant to take cold calls, share personal data or spend time filling in surveys. The pandemic also produced a lot of statistical "noise" For the ONS it was acute. Analysis of other sources showed at one point the ONS was underestimating the size of Britain's workforce by as much as one million people. Stats for certain sub-groups were swinging by 30 per cent in a month because they depended on a sample size of five. In October 2023, the ONS was forced to temporarily pause its monthly readout on jobs
A picture of Keir Starmer, Rachel Reeves and some other people in a hospital. They look very odd.
Quote from Llama paper: "During training, tens of thousands of GPUs may increase or decrease power consumption at the same time, for example, due to all GPUs waiting for checkpointing or collective communications to finish, or the startup or shutdown of the entire training job. When this happens, it can result in instant fluctuations of power consumption across the datacenter on the order of tens of megawatts, stretching the limits of the power grid. This is an ongoing challenge for us as we scale training for future, even larger Llama"
Gemini CLI. First there was Claude Code in February, then OpenAl Codex (CLI) in April, and now Gemini CLI in June. All three of the largest Al labs now have their own version of what I am calling a "terminal agent" - a CLI tool that can read and write files and execute commands on your behalf in the
It was noted by many individuals that the Principal, who was an excellent public speaker, tended to control the narrative to Court and elsewhere, consistently painting a picture that all was well in the University, highlighting the institution's successes, working well with donors and external stakeholders, rightly praising the achievements of students and bestowing esteem upon the University by his election as Convenor of Universities Scotland. There were however many reports to this investigation that the Principal frequently demonstrated hubris, or excessive pride in his role. Hubris is relevant insofar as it can be characterised by dangerous over self-confidence and complacency, often in combination with arrogance, contempt towards people who offer criticism and obsession with personal image and status. Additional comments were noted regarding the Principal's overbearing leadership style, behaviours and dislike of potentially awkward confrontations or questioning and the potential adverse impact these factors may have had upon individuals and the overall culture of engagement within the Universi
We have seen no record that universities that had already implemented Blueprints were asked by the University about their experience of using it. In June 2023 the Admissions team raised major concerns regarding the amount of re-work needed to their processes. In February 2024, the project risk status was changed to 'red' • In July 2024 a list of 'configuration items' (changes that the University wanted) was shared with Tribal who were unable to give target dates for their implementation. Despite this, a go-live decision was made for September 2024. Following a short delay the system did go live but, according to the University, was unable to be used for seven weeks. No 'work arounds' for offers or admissions had been developed in advance of the system going live and, as the data had been migrated from the old system to Blueprints, we were told that it was not possible to introduce a manual system at the time or revert to the old system which had been 'switched off'
During the period from 1 January 2023 to 31 July 2024, cash balances decreased from c.£ 100m to c. £32m. At the same time, there was a lack of consistency around the reporting of cash and liquidity issues to UEG, FPC and Court. Consequently, no action was taken during this period to arrest the decline, and, at the time of this Investigation, the University is forecasting (as of May 2025) that the closing cash balance at the end of FY25 will be £1m overdrawn.
Photo showing the Satchvia band taking their applause at the end of the gig
Allow me also to make the point again that this doesn't absolve us from thinking deeply about the coming explosion due to energy needs. Here as well, I am sticking to my guns, arguing that we may currently be underestimating our overall energy needs, and dramatically so. In short: While a concern about exploding energy requirements in a world full of Al agents remains entirely defensible, regular chatbot use is really nothing to worry about in terms of energy use.
"In 2025, we find that the student use of Al has surged in the last year, with almost all students (92%) now using Al in some form, up from 66% in 2024, and some 88% having used GenAl for assessments, up from 53% in 2024. The main uses of GenAl are explaining concepts, summarising articles and suggesting research ideas but a significant number of students - 18% - have included Al-generated text directly in their work."
"The email I'd have written is actually shorter than the original prompt, which means I spent more time asking Gemini for help than I would have if l'd just written the draft myself. Remarkably, the Gmail team has shipped a product that perfectly captures the experience of managing an underperforming employee"