Reuben Binns

What do LLMs see?

I wrote a lil' tool that extracts the attention matrices out of open models and creates this typing visual, with each token's opacity changing according to its average attention score as the prompt progresses. Dimmer words are considered less important to the model.

Despite working 60+ hour weeks, many Uber and Lyft drivers have seen their pay drop by nearly 50% over time. Even as these companies report record profits, consumers face inflated prices, while the workers who are (quite literally) driving these companies’ success are paid less and less.

Grok Exposed a Porn Performer’s Legal Name and Birthdate—Without Even Being Asked In the latest in a string of privacy abuses from the chatbot, Grok provided porn performer Siri Dahl's full legal name and birthdate to the public, information she'd protected until now.

In the latest in a string of privacy abuses from the chatbot, Grok provided porn performer Siri Dahl's full legal name and birthdate to the public, information she'd protected until now.

a windows 11 menu showing "convert to adobe pdf", "ask copilot", and 4 more options that take time to load

windows 11 update: for the first time in my life, simply right clicking a file has a load time. unreal

If you were to create a philosophy of LLM course, what kinds of things would you cover?

My thoughts in thread, but eager to hear what others would do

The key finding is that switching from a chronological to an algorithmic feed shifted users toward more conservative positions: policy priorities, views on Trump’s criminal investigations, and attitudes toward the war in Ukraine all shifted to the right.

I hacked ChatGPT and Google's AI – and it only took 20 minutes I found a way to make AI tell you lies – and I'm not the only one.

I got a tip that all over the world, people are using a dead-simple hack to manipulate AI behavior. It turns out changing what AI tells other people can be as easy as writing a blog post *on your own website*

I didn’t believe it, so I decided to test it myself www.bbc.com/future/artic...

My decade long struggle to achieve AI aligned to my human values: autocorrect still won't let me type duck. Duck. Ah for duck's sake

This polling suggests the possibility of a party (Reform) winning most votes, despite also being a Condorcet loser - i.e. a party who would lose in a head-to-head against any other party. Unprecedented, but possible under first past the post - almost impossible under other electoral systems.

Sure:
Light rail, high speed rail, ocean wave generation, solar power generation, wind generation, heat pump tech, geothermal heat pump tech, last mile on-demand transit solutions, better wheelchairs, better wheel chair lifts, better flexible seating on public transit, maker spaces, tool libraries.

I had the pleasure of being a witness at the Joint Committee on Artificial Intelligence of the @oireachtasnews.bsky.social this morning, to talk about AI, truth and democracy: www.oireachtas.ie/en/press-cen...

There is speculation that the introduction of driverless cars could lead to jaywalking laws in the UK

65% of Britons would oppose making jaywalking a minor offence, including 41% "strongly opposed"

Only 24% are would support such rules

yougov.co.uk/topics/trave...

Banner image with screenshot of scientific article from nature Medicine, as well as two panels from the study method and results

⚠️ Despite all the hype, chatbots still make terrible doctors. Out today is the largest user study of language models for medical self-diagnosis. We found that chatbots provide inaccurate and inconsistent answers, and that people are better off using online searches or their own judgment.

Again, in another context - "human-free code"

bsky.app/profile/simo...

the human agents were only there to “provide guidance” to the robotaxis. “They do not remotely drive the vehicles. Waymo asks for guidance in certain situations and gets human input. But the Waymo vehicle is always in charge of the dynamic driving task,” Peña said.

A total inversion of the old 'human in the loop' defense:

"Waymo asks for guidance in certain situations and gets human input. But the Waymo vehicle is always in charge of the dynamic driving task"

cybernews.com/news/waymo-o...

Please, my magic bean, it is very sick

Please consider applying for the 10th edition of CNIL-Inria Privacy Award.

The CNIL-Inria Privacy Award is intended to promote research in Computer Science and Privacy and to raise awareness among citizens and decision-makers on privacy and data protection issues.

1/2

If only computers had built in clocks in order to tell time using some sort of deterministic machine-readable format.

But alas, the only way to know whether it's daytime is to pay nearly a dollar every 30 minutes to ask a fancy autocomplete bot.

I wonder if readers of published papers will also be informed of which policy applied to which paper

LLM Policy Preference* ICML 2026 will follow a two-policy framework for the use of large language models (LLMs) in reviewing, based on the following two policies: Policy A (Conservative): Use of LLMs for reviewing is strictly prohibited. Policy B (Permissive): Allowed: Use of LLMs to help understand the paper and related works, and polish reviews. Submissions can be fed to privacy-compliant* LLMs. Not allowed: Ask LLMs about strengths/weaknesses, ask to suggest key points for the review, suggest an outline for the review, or write the full review. *By “privacy-compliant”, we refer to LLM tools that do not use logged data for training and that place limits on data retention. This includes enterprise/institutional subscriptions to LLM APIs, consumer subscriptions with an explicit opt-out from training, and self-hosted LLMs. (We understand that this is an oversimplification.) Under both policies, reviewers are always responsible for the full content of their reviews. Both authors and reviewers declare their preferences, and submissions are matched with compatible reviewers. Each reviewer is told which policy to follow on all of their assigned papers (we try to minimize the number of overrides of reviewer preferences). Any reviewer who is an author on a paper that requires Policy A must also be willing to follow Policy A. Reviewing data will be analyzed to inform future policies. For full details, see https://icml.cc/Conferences/2026/LLM-Policy. Please indicate which policy you prefer as a reviewer. Keep in mind that if you are an author on a paper that requires Policy A, you must be willing to follow Policy A as a reviewer. I strongly prefer Policy A. I am okay with either A or B. I strongly prefer Policy B.

For ICML 2026 (top tier AI research venue), authors can choose between two policies for the use of LLMs in reviewing:

Policy A: LLMs prohibited (authors who chose this are bound by it when reviewing)

Policy B: "Use of LLMs to help understand the paper and related works, and polish reviews" allowed

Edinburgh City Councillors meet food delivery workers in Workers' Observatory workshop

🎯 Getting Gig Work on Scotland's Political Agenda

Food delivery workers joined the Workers' Observatory workshop this week to build political momentum towards fair and safe gig work in Scotland.

You can read the workers' principles and policy demands here: workersobservatory.org/documents/WO...

thanks ana! actually, i think these particular demonstrations are featuring original Enigmas, and delivered by the Bletchley Park folk, not my home-brewed version (the photo is confusing!)

Matt Goodwin's lies about his background have hopelessly confused Google AI 😉

Stack of the book ‘In Solidarity, Under Suspicion: The British Far Left from 1956’, edited by Daniel Frost and Evan Smith

To celebrate the publication of @d-j-frost.bsky.social and mine’s book ‘In Solidarity, Under Suspicion: The British Far Left from 1956’, I am giving away a copy to one randomly chosen person who reposts this post by 11.59pm ACST Sunday 1 Feb.

manchesteruniversitypress.co.uk/9781526179593/

For 1), see independenttechresearch.org

And consider joining - strength in numbers!

Main Responsibilities Operational and Financial Administration: • Coordinate the day-to-day operational and administrative activities of the AI Accountability Lab. • Oversee lab budgets and financial planning for research grants and cost centres; monitor expenditure and prepare financial reports for the PI and Faculty Finance Office. • Ensure adherence to Trinity and funder financial policies and procurement procedures. • Maintain internal administrative systems, documentation, and records to support efficient project delivery. • Liaise with central Finance, the Research Development Office, and external funding bodies on budgetary and governance matters. • Coordinate HR administrative processes for the lab, including recruitment, onboarding, contract administration, and research staff extensions in liaison with central HR and the Research Office. • Create and maintain shared resources, calendars, and internal documentation. • Coordinate relationships and collaboration with other research centres, government bodies, policy makers, civil society partners, and media as required by the research team. • Plan and schedule internal lab meetings, reading groups, invited speakers, collaborative work sessions, and other events as required.

Research Coordination and Governance: • Track project timelines, deliverables, reporting schedules, and ethical compliance across multiple funded projects. • Ensure research activities meet Trinity and funder governance requirements (e.g., ethics, data management, GDPR, and reporting). • Coordinate workshops, seminars, and collaborative events with research centres and external partners as required. • Maintain shared documentation and project records to support effective collaboration. • Ensure that research outputs comply with ethical and scientific standards and satisfy the terms and conditions of relevant funding bodies.

Communications and External Engagement • Coordinate internal and external communications for the lab, including its website, newsletters, and social media presence. • Liaise with policy makers, media, and stakeholder organisations on behalf of the lab, in consultation with the PI. • Support the dissemination of research outputs, such as academic publications, policy briefs, and public events in collaboration with ADAPT’s and TCD’s communications staff. • Work with Trinity Communications to prepare press releases and highlight lab achievements. • Representing the lab externally with media if required. Planning and Cross-Centre Coordination • Support the PI in delivering AIAL’s annual operational plan and reporting on progress against objectives. • Coordinate activities across multiple projects and partnerships to ensure consistency and alignment with institutional priorities. • Contribute to process improvements that strengthen the lab’s administrative and operational framework. • Act as a key contact for internal governance reviews, audits, and funder compliance checks. Any other duties or responsibilities as assigned by the PI or their delegate in support of the effective operation of the AI Accountability Lab and its objectives.

I’m looking for my right-hand person to come help me run the @aial.ie

- Job Title: Lab Coordinator, AI Accountability Lab (0.8 FTE)
- Pay Scale: (€58,999 - €69,325 per annum pro-rata)
- Closing Date: 11-Feb-2026 12:00

Apply here: my.corehr.com/pls/trrecrui...

Main Responsibilities👇🏾

Back to GDP:

There is no productivity miracle in the US.

The growth of productivity (= GDP per hour worked) has slowed down markedly.

From 1.9% in the period 1980-2010 to 1.1% in 2010-2025

If you think the rise of tech has made people more productive, think again

Meanwhile, moving tuition overseas means the UK economy loses out on student spending (accommodation, living expenses), staff wages, estates+facilities. Not to mention all the multiplier effects for local economies around universities.

UK ministers scrap foreign students target in shift to overseas hubs strategy Government replaces recruitment goal with plan to increase ‘education exports’ to £40bn a year by 2030

This is being pitched as a 'plan to increase ‘education exports’', but fees from international students at UK institutions are already considered exports (under WTO definitions, "Mode 2: Consumption Abroad").

www.theguardian.com/politics/202...