Resource Hub
3505 hand-picked resources for students, job seekers, career changers, founders, and AI learners — 100% free to browse with no account required.
Filtered by tag
Answers come only from resources in this hub, with the sources listed.
267 resources
Kevin Roose's newsletter post that publishes Dario Amodei's 2017 internal OpenAI memo, "The Big Blob of Compute Hypothesis," in full for the first time, with background on how it argued for making models bigger rather than more specialized and how it shaped OpenAI's early direction. Free to read on Substack.
Free-to-read online essays and interviews about science, technology, economics and culture. Asterisk’s about page describes a San Francisco Bay Area perspective and a focus on examining assumptions and evidence. Published by nonprofit Obelus, Inc.; supported by donors, subscriptions and a Coefficient Giving grant. Print subscriptions cost money; this entry links to free online reading.
JFF's explainer on three-year bachelor's degree programs: how schools are redesigning degrees around fewer credits and clearer career outcomes, which accreditors have approved them, and what the model could mean for cost and time-to-degree.
Why I recommend it: Worth reading before you or someone you advise commits to a four-year price tag. JFF is a workforce nonprofit with its own agenda around shorter pathways, so treat the optimism as advocacy — but the accreditation facts and school examples check out.
A communal blog where working mathematicians — professors, PhD students, sceptics and enthusiasts alike — write about what AI is doing to their field: authorship, what counts as understanding, where papers will go, and whether the job changes. Recent pieces include Martin Hairer on why he joined the AGMAI advisory group and Jonny Evans on the choices ahead. Submissions are open to anyone in the field, any length.
Why I recommend it: Free to read and free to write for. Worth reading even if you never touch mathematics: it is one of the few places where a whole profession is arguing in public about what AI does to its craft, in its own words rather than a journalist's. These are individual opinions, not findings — the value is the range of them, and the disagreement is the point.
Free reading app from OverDrive that borrows ebooks, audiobooks and magazines from your local public library with a library card. Works in a browser or on iOS, Android, Kobo and Kindle (US libraries only), with offline downloads, CarPlay and Android Auto support, holds, and tagged reading lists.
Why I recommend it: This is how you read the paid books on this site without buying them — most business, career and technology titles are in public library collections. What you need is a library card, which is free where you live. The catch is availability, not price: your library chooses what it licenses and popular titles come with waiting lists, so place holds early rather than expecting a book on the day you want it.
The free research library and blog of Snorkel AI, the company spun out of Stanford's Snorkel project on programmatic labelling. The papers and posts explain how training data for AI models is actually built — labelling, evaluation sets, and the 'environments' used to train agents. Useful if you want to understand the unglamorous data work behind model quality, which is where a lot of the real jobs are.
Why I recommend it: Research and blog posts are free to read with no signup. Read it for the how, not the verdict: Snorkel sells data services to frontier AI labs, so posts arguing that better data beats bigger models are also a sales case. Everything else on the site is a paid enterprise product — 'request dataset samples' means a sales call.
Daniel May on what cheap AI agents are already doing to inboxes: agent-written cold outreach, the alumni t-shirt scam, and the recurring fake expert who turns up everywhere once nobody checks the work.
Why I recommend it: Free to read. Worth twenty minutes if you send or receive cold messages: it shows exactly what automated outreach looks like from the other side, and why a message that reads fluently now proves nothing about whether a person wrote it.
Turns books and documents into scrollable, bite-sized learning cards with AI summaries and quizzes.
From the site: What is AI smart scrolling? ScrollEd pioneered AI smart-scrolling technology that transforms books into scrollable, bite-sized learning cards with AI summaries and quizzes. Learn faster in 5-minute sessions.
Why I recommend it: The free tier is limited: demo content only, no file uploads, and ads. Pro is $6.99 a month (currently offered free for three months to early members), so treat the free plan as a look around rather than a usable reading tool.
Raji and colleagues examine the ethics of the audits themselves: who is photographed, who consented, and what an auditor owes the people in the test set.
From the site: Although essential to revealing biased performance, well intentioned attempts at algorithmic auditing can have effects that may harm the very populations these measures are meant to protect. This concern is even more salient while auditing biometric systems such as facial recognition, where the data is sensitive and t…
Why I recommend it: A rare paper about the ethics of doing ethics work. Free on arXiv.
Vinay Prabhu and Abeba Birhane examine widely used image datasets and find non-consensual photos of real people, offensive labels and no realistic route to consent.
From the site: In this paper we investigate problematic practices and consequences of large scale vision datasets. We examine broad issues such as the question of consent and justice as well as specific concerns such as the inclusion of verifiably pornographic images in datasets. Taking the ImageNet-ILSVRC-2012 dataset as an example…
Why I recommend it: This is the audit that got a major benchmark dataset withdrawn. Short, readable, and free.
Birhane and colleagues show that training on a larger scrape makes hateful content and racist misclassification worse, not better — direct evidence against "more data fixes it".
From the site: `Scale the model, scale the data, scale the GPU-farms' is the reigning sentiment in the world of generative AI today. While model scaling has been extensively studied, data scaling and its downstream impacts remain under explored. This is especially of critical importance in the context of visio-linguistic datasets wh…
Why I recommend it: Useful whenever someone argues scale solves bias. Free in full on arXiv.
Abeba Birhane, Vinay Prabhu and Emmanuel Kahembwe audit the LAION-400M dataset used to train popular image models and document the racist, misogynistic and non-consensual material inside it.
From the site: We have now entered the era of trillion parameter machine learning models trained on billion-sized datasets scraped from the internet. The rise of these gargantuan datasets has given rise to formidable bodies of critical work that has called for caution while generating these large datasets. These address concerns sur…
Why I recommend it: The paper to read before anyone tells you a model is fine because the data was "publicly available". Free in full on arXiv.
Raji and colleagues argue many deployed AI systems fail on their own stated terms — they simply do not work — and that this belongs in the harm conversation alongside bias.
From the site: Deployed AI systems often do not work. They can be constructed haphazardly, deployed indiscriminately, and promoted deceptively. However, despite this reality, scholars, the press, and policymakers pay too little attention to functionality. This leads to technical and policy solutions focused on "ethical" or value-ali…
Why I recommend it: The first question is not "is it fair" but "does it work at all". Free in full on arXiv.
Raji and co-authors show that benchmarks claiming to measure general ability measure something much narrower, and that the gap is how overclaiming happens.
From the site: There is a tendency across different subfields in AI to valorize a small collection of influential benchmarks. These benchmarks operate as stand-ins for a range of anointed common problems that are frequently framed as foundational milestones on the path towards flexible and generalizable AI systems. State-of-the-art…
Why I recommend it: Read this before you trust a benchmark chart in a launch post. Free on arXiv.
Deborah Raji and co-authors set out a practical, stage-by-stage internal audit process for AI systems, from scoping through to post-deployment review.
From the site: Rising concern for the societal implications of artificial intelligence systems has inspired a wave of academic and journalistic literature in which deployed systems are audited for harm by investigators from outside the organizations deploying the algorithms. However, it remains challenging for practitioners to ident…
Why I recommend it: The closest thing to a step-by-step audit template you can use inside an organization. Free on arXiv.
Five plain-language protections people should expect from automated systems — safe systems, protection from discrimination, data privacy, notice and explanation, and a human alternative — each with a section on what organisations should do.
From the site: Among the great challenges posed to democracy today is the use of technology, data, and automated systems in ways that threaten the rights of the American public. Too often, these tools are used to limit our opportunities and prevent our access to critical resources or services. These problems are well documented. In…
Why I recommend it: Led by Alondra Nelson at the White House science office in 2022. It was never binding and has since been removed from whitehouse.gov, so this link goes to the official archive — still the clearest short statement of what people are owed.
Arvind Narayanan and Sayash Kapoor's newsletter separating AI that works from AI that is oversold, with close readings of specific product and research claims.
From the site: Analyzing AI as transformative but normal technology, not superintelligence. Click to read AI as Normal Technology, a Substack publication with tens of thousands of subscribers.
Why I recommend it: The archive is free to read and the best regular antidote to launch-post hype. The book of the same name is a paid purchase — you do not need it to follow the newsletter.
Proposes a short standard document to ship with every trained model: what it is for, who it was tested on, where it performs worse, and what it should not be used for. Most model documentation you see today descends from this. Free on arXiv.
From the site: Trained machine learning models are increasingly used to perform high-impact tasks in areas such as law enforcement, medicine, education, and employment. In order to clarify the intended use cases of machine learning models and minimize their usage in contexts for which they are not well suited, we recommend that rele…
Why I recommend it: This is the practical end of AI ethics — a template, not an argument. Useful if you ever have to evaluate a vendor's model.
Separates the technical question (how do you make a system pursue a goal) from the normative one (whose values, chosen how). Argues no single person's values are a legitimate target and looks at fair-process alternatives. Free on arXiv.
From the site: This paper looks at philosophical questions that arise in the context of AI alignment. It defends three propositions. First, normative and technical aspects of the AI alignment problem are interrelated, creating space for productive engagement between people working in both domains. Second, it is important to be clear…
Why I recommend it: The clearest philosophical treatment of “aligned to what?” I have found. Skip the lab blog posts and read this instead.
A long report on what changes when AI acts on your behalf rather than answering questions: manipulation, anthropomorphism, misaligned delegation, and what happens when everyone has an assistant at once. Free on arXiv.
From the site: This paper focuses on the opportunities and the ethical and societal risks posed by advanced AI assistants. We define advanced AI assistants as artificial agents with natural language interfaces, whose function is to plan and execute sequences of actions on behalf of a user, across one or more domains, in line with th…
Why I recommend it: Dense but the most thorough thing published on agent ethics. Use the section headings to read only the parts you need.
Argues that before release, labs should test models for dangerous capabilities and for whether they will apply them — and sets out what responsible release decisions would look like. Free on arXiv.
From the site: Current approaches to building general-purpose AI systems tend to produce systems with both beneficial and harmful capabilities. Further progress in AI development could lead to capabilities that pose extreme risks, such as offensive cyber capabilities or strong manipulation skills. We explain why model evaluation is…
Why I recommend it: This is where today's “frontier safety framework” language comes from. Written largely by the labs it would govern, which is worth holding in mind.
Argues that ever-larger language models carry costs that scale with them: environmental cost, unauditable training data, encoded bias, and the illusion of understanding. The source of the phrase “stochastic parrot.” Free to read on the ACM site.
Why I recommend it: The paper that got two of its authors pushed out of Google. Worth reading before you accept either the hype or the dismissal of it.
Borrows the electronics-industry datasheet idea for training data: how it was collected, who is in it, who consented, and what it should not be used for. Free on arXiv.
From the site: The machine learning community currently has no standardized process for documenting datasets, which can lead to severe consequences in high-stakes domains. To address this gap, we propose datasheets for datasets. In the electronics industry, every component, no matter how simple or complex, is accompanied with a data…
Why I recommend it: Pairs directly with Model Cards. Together they are the closest thing the field has to a documentation standard.
Reviewed 146 papers on bias in language technology and found most never say who is harmed or how. Argues bias work has to start from real-world power relations, not just from a metric. Free on arXiv.
From the site: We survey 146 papers analyzing "bias" in NLP systems, finding that their motivations are often vague, inconsistent, and lacking in normative reasoning, despite the fact that analyzing "bias" is an inherently normative process. We further find that these papers' proposed quantitative techniques for measuring or mitigat…
Why I recommend it: Read this if “bias” has started to sound like a box to tick. It is a careful takedown of shallow fairness work by people who do the work.
