Adam Jones's Blog

new: Announcing... AdamCon 2026! (published 5 days ago)

selected

on AI safety

on everything else

all

There are a lot of posts here, on quite different topics. This site is fairly friendly for AI agents to browse, so one way to find what's relevant to you is to paste this into your AI assistant of choice:

Browse https://adamjones.me/blog/ and the posts linked from it. Based on what you know about me, suggest 3–5 posts that are most relevant to me, with a sentence on why for each.

Announcing... AdamCon 2026! — 21 June 2026
A guide to building useful AI tools for AI safety — 8 June 2026
Good MCP servers are designed for composition — 7 June 2026
I'm hiring: Software Engineer, RL Data at Anthropic — 3 June 2026
How we contain Claude across products External — 25 May 2026
Local, real-time text-to-speech: compare Supertonic 2's 10 voices — 17 April 2026
Running LLMs Locally in 2026: Local models got ~2.8x faster in a year — 16 April 2026
What's the best dictation app? I tried 21 Wispr Flow alternatives to find out. — 15 April 2026
How to set up the Slack MCP server with a custom client — 14 April 2026
Working at Anthropic FAQs — 11 April 2026
What is the Low Calorie Meal (LCML) on United Airlines? — 7 April 2026
How a coalition of democracies solves AGI's biggest risks — 21 March 2026
UBI doesn't solve AI power concentration — 15 March 2026
Open source doesn't solve AI power concentration — 15 March 2026
How AGI could kill US democracy — 15 March 2026
Give AI companies something to aim for: The case for beneficial capability evals External — 10 January 2026
Langfuse Context: All things MCP with Adam Jones (Tech Lead at Anthropic) External — 6 January 2026
MCP Connect with Anthropic, Zoopla & Alpic | London edition External — 19 December 2025
Code Execution with MCP: Fix Tool Token Bloat (Adam Jones, Anthropic) External — 12 December 2025
Code execution with MCP: Building more efficient agents External — 4 November 2025
The Model Context Protocol: Connecting AI to Everything - Adam Jones, Anthropic External — 3 November 2025
An AI safety plan that might work: an international coalition governing AGI — 15 September 2025
What should the UK be doing to make AGI go well? — 15 August 2025
Announcing... AdamCon 2025! — 18 July 2025
Overcoming problems with compute thresholds for AI regulation — 30 June 2025
AI Model Thresholds: How Governments Can Identify Frontier AI Systems — 30 June 2025
5 ways to develop an AI safety plan or field strategy External — 29 June 2025
Quantitative models of AI-driven bioterrorism and lab leak biorisk — 25 June 2025
What makes a good AI safety plan? External — 20 June 2025
Key paths, plans and strategies to AI safety success External — 19 June 2025
How you can get more value from me reviewing your work — 18 June 2025
Preferences for communicating with me — 18 June 2025
LinkedIn Ads: can you get more efficient marketing by overbudgeting and holding your manual bid the same? External — 17 June 2025
What is customer due diligence in AI safety? External — 13 June 2025
Export ALL Your WhatsApp Chats from Android to Your Computer! External — 25 May 2025
What might AGI look like, concretely? — 21 May 2025
Why I disagree with Yann LeCun on whether LLMs could scale to AGI — 11 May 2025
300,000 people are directly creating training data for AI — 21 April 2025
What does Yann LeCun think about AGI? A summary of his talk, "Mathematical Obstacles on the Way to Human-Level AI" — 5 April 2025
Major UK banks are training their customers to fall for scams — 4 April 2025
Running LLMs Locally in 2025: Speed tests on M2 Pro + 16 GB RAM — 2 April 2025
A rough plan for AI alignment assuming short timelines — 31 March 2025
YouTube series: How to contribute to the BlueDot Impact repo External — 30 March 2025
How to set up PostHog for a Bubble single-page application, with proper pageview tracking — 24 March 2025
AI safety content you could create — 6 January 2025
Policymakers don't have access to paywalled articles — 5 January 2025
Alignment Is Not All You Need: Other Problems in AI Safety — 2 January 2025
The post-AGI purpose problem — 1 January 2025
Why product managers are uniquely suited for tech policy roles — 31 December 2024
The beginners guide to investing (2025 UK edition) — 25 December 2024
Setting up OpenWrt on the DSL-AC68U for 1 Gig speeds — 22 December 2024
Teach-swap-explain: a learning activity for course designers to create highly effective learning experiences External — 12 December 2024
Why we run our AI safety courses External — 4 December 2024
How Does AI Learn? A Beginner’s Guide with Examples External — 29 November 2024
The standard W3C Gamepad API mapping for an Xbox controller — 4 November 2024
What early career policymakers can learn from product managers: understanding people’s actual problems is key to effective policy — 1 November 2024
AI Alignment June 2024 course retrospective External — 30 October 2024
OpenAI's cybersecurity is probably regulated by NIS Regulations — 25 October 2024
Does project proposal feedback result in better final projects? External — 21 October 2024
No time for user interviews? Learn how to use empathetic role-playing to make better product decisions. — 2 October 2024
An easy win for UK AI safety: competition law safe harbour — 23 September 2024
Modular AI Safety courses proposal External — 16 September 2024
Summary of AI alignment participant user interviews External — 13 September 2024
An easy win for UK AI safety: supporting whistleblowers — 11 September 2024
What we didn’t cover in our June 2024 AI Alignment course (or, an accessible list of more niche alignment research agendas) External — 13 August 2024
The AI regulator's toolbox: A list of concrete AI governance practices — 10 August 2024
Are cheap shaver blades any good? — 8 August 2024
Advertising to technical people: LinkedIn, Twitter, Reddit and others compared External — 7 August 2024
What advertising creatives work for technical people? External — 7 August 2024
Results from testing ad adjustments External — 7 August 2024
Diagnosing infectious diseases with CRISPR: SHERLOCK and DETECTR explained — 1 August 2024
Reflections on my 7-day writing challenge — 23 July 2024
How to avoid the 2 mistakes behind 89% of rejected AI alignment applications External — 22 July 2024
What do applicants mean when they say they come from LinkedIn? External — 19 July 2024
Our 2023 internal cybersecurity course External — 15 July 2024
Addressing digital harms: a right of appeal is not sufficient — 15 July 2024
AI as a corporation (or, an intro to AI safety?) — 14 July 2024
How to fix proof of address — 13 July 2024
Proof of address is nonsense — 12 July 2024
Government departments should say they don't care — 11 July 2024
Avoiding unhelpful work as a new AI governance researcher — 9 July 2024
Asking me for help — 9 July 2024
Preventing overreliance: The case for deliberate AI errors in human-in-the-loop systems — 8 July 2024
Why having a human-in-the-loop doesn't solve everything — 7 July 2024
7 blogs in 7 days — 7 July 2024
What we learnt from running our AI alignment course in March 2024 External — 2 July 2024
What is a lead cohort? External — 1 July 2024
What we changed for the June 2024 AI alignment course External — 24 June 2024
A thing I'd like to exist: benchmarks for train internet — 23 June 2024
3 articles on AI safety we’d like to exist External — 17 June 2024
Why we work in public at BlueDot Impact External — 16 June 2024
Why are people building AI systems? External — 25 May 2024
How to send Keycloak emails through Google Workspace's SMTP relay — 18 May 2024
Preventing AI oligarchies: important, neglected and tractable work — 12 May 2024
How might AI-enabled oligarchies arise? — 11 May 2024
Follow-up: benchmarking Next.js server vs nginx at serving a static site, now on AWS — 30 April 2024
Benchmarking the Next.js server vs nginx at serving a static site — 29 April 2024
No, I don’t want to fill out your contact form — 28 April 2024
AI alignment project ideas External — 12 April 2024
How to avoid the 4 mistakes behind 92% of rejected AI governance applications External — 5 April 2024
Do cheap GPS trackers work? A review of the GF-07, GF-09 and GF-22. — 24 March 2024
Can we scale human feedback for complex AI tasks? An intro to scalable oversight. External — 18 March 2024
ai-safety.txt: Making AI vulnerability reporting easy — 17 March 2024
What's the best Myprotein flavour? I tried 23 of them to find out. — 10 March 2024
OHGOOD: A coordination body for compute governance — 3 March 2024
What is AI alignment? External — 1 March 2024
What risks does AI pose? External — 21 February 2024
How are AI companies doing with their voluntary commitments on vulnerability reporting? — 20 February 2024