# Research Briefs ---- ## Overview | | | | ----------------------------------------------------------------------------------- | ------------------------------------------------------ | | **[[#Flagship brief my perspective on the AI landscape\|Flagship]]** | **My perspective on the AI landscape** | | **[[#Analysis briefs evidence dissected and interpreted\|Analysis]]**   | **Evidence dissected and interpreted**      | | **[[#Clarity briefs key technologies, concepts, and research explained\|Clarity]]** | **Key technologies, concepts, and research explained** | | **[[#Spark briefs food for thought, some of it surprising\|Sparks]]** | **Food for thought, some of it surprising** | | **[[#Conjecture briefs grounded speculations\|Conjectures]]** | **Grounded speculations** | | **[[#Source briefs external material other briefs refer to\|Sources]]** | **External material other briefs refer to** | ### Forthcoming briefs To hear about new research when it goes live, see [[Subscribe|Research > Subscribe]]. --- ## Flagship brief: my perspective on the AI landscape > **[[Industry and Society Need AI Clarity|Industry and Society Need AI Clarity]]** > **The opportunity in today’s confused and divided AI landscape** --- ## Analysis briefs: evidence dissected and interpreted > [[Briefs#Forthcoming briefs|Coding Agent Adoption (forthcoming)]] > Quantifying the rapid spread of agentic coding > [[Briefs#Forthcoming briefs|Coding Agent ROI (forthcoming)]] > The benefits of coding agents are uneven — so far > [[Briefs#Forthcoming briefs|Coding and Agents Dovetail (forthcoming)]] > The nature of software development makes it a strong fit for agentic AI > [[Briefs#Forthcoming briefs|Agentic Coding’s Impact on Agentic AI (forthcoming)]] > Enterprises will reshape other kinds of work to more closely resemble coding > [[Briefs#Forthcoming briefs|The Future of AI MechInterp (forthcoming)]] > How mechanistic interpretability’s early successes will play out --- ## Clarity briefs: key technologies, concepts, and research explained > **[[Why Care About Clarity on AI|Why Care About Clarity on AI]]** > **Better understanding of AI is key to steering it toward positive outcomes** > [[Briefs#Forthcoming briefs|Clarity on AI Assistants (forthcoming)]] > Understand how ChatGPT, Claude, and others do what they do > [[Briefs#Forthcoming briefs|Clarity on AI Model Training (forthcoming)]] > Understand what, how, and when AI models learn > [[Briefs#Forthcoming briefs|Clarity on AI Interpretability (forthcoming)]] > Peering into how Al models seem to think — and why it matters > [[Briefs#Forthcoming briefs|Clarity on AI Safety (forthcoming)]] > Ensuring AI acts in accordance with human goals and values > [[Briefs#Forthcoming briefs|Clarity on AI Alignment (forthcoming)]] > Ensuring AI acts in accordance with human goals and values > [[Briefs#Forthcoming briefs|Clarity on AI Autonomy (forthcoming)]] > When to let AI systems act autonomously — and for how long > [[Briefs#Forthcoming briefs|Clarity on AI Reasoning (forthcoming)]] > The various ways AI systems seem to reason — and when to trust them > [[Briefs#Forthcoming briefs|Clarity on AI Epistemology (forthcoming)]] > What AI systems seem to believe and why — and when to trust them > [[Briefs#Forthcoming briefs|Clarity on AI Functionalism (forthcoming)]] > AI consciousness, why it mostly doesn’t matter — and when it might > [[Briefs#Forthcoming briefs|Clarity on AI Model Welfare (forthcoming)]] > Considering whether AI systems experience anything — and the implications --- ## Spark briefs: food for thought, some of it surprising > **[[From Shakespeare to AI|From Shakespeare to AI]]** > **How LLMs’ inner workings are rooted in comparative literature** > [[Briefs#Forthcoming briefs|Transparent Black Boxes (forthcoming)]] > Neural networks are well understood — just unfathomable at large scale > [[Briefs#Forthcoming briefs|The Autocomplete Fallacy (forthcoming)]] > Why predicting what comes next is more powerful than it seems it should be > [[Briefs#Forthcoming briefs|We Are ChatGPT (forthcoming)]] > Who you’re talking with when you talk with an LLM-powered AI assistant > [[Briefs#Forthcoming briefs|Hallucination Is Forever (forthcoming)]] > Inaccurate statements from language models can be contained but never prevented > [[Briefs#Forthcoming briefs|AI Isn’t Biased (forthcoming)]] > Bias is caused by choices of training data — it’s not in AI algorithms > [[Briefs#Forthcoming briefs|When Your Mind Autocompletes (forthcoming)]] > Essentials of large language models explained with a thought experiment --- ## Conjecture briefs: grounded speculations > [[Briefs#Forthcoming briefs|The Fate of Work (forthcoming)]] > What people will do with their time if AI obviates the need for human work > [[Briefs#Forthcoming briefs|Animal AI (forthcoming)]] > When a machine emulates the cognition of an octopus > [[Briefs#Forthcoming briefs|When AI-Free Means Luxury (forthcoming)]] > Some “human-made” work will become more valuable once AI is everywhere --- ## Source briefs: external material other briefs refer to > **[[The Polarized Predictions About AI|The Polarized Predictions About AI]]** > **Influential viewpoints range from utopian to dystopian** > **[[LLMs Don't Run on Facts or Logic|LLMs Don't Run on Facts or Logic]]** > **Leading AI researchers explain the limitations of large language models** > [[Briefs#Forthcoming briefs|The Many Definitions of AI (forthcoming)]] > There’s Widespread Disagreement About What “AI” Means > [[Briefs#Forthcoming briefs|The Many Definitions of AI Agents (forthcoming)]] > There’s Widespread Disagreement About What “AI Agent” Means --- ## Suggestions? If you’d like to suggest an AI-related question, topic, or tension you think deserves a research brief, [[Contact|reach out]]. ## About my research Learn more about: - My research [[Methods|methods]] - The [[Scope|scope]] of my research - [[Research/Sharing|Sharing]] (or otherwise reusing) my research - How I use — and don’t use — [[Tools|AI as a tool]] in my research process ----