LessWrong (30+ Karma)
Episodes
“Phantom Transfer and the Basic Science of Data Poisoning” by draganover
15 Feb 2026
Contributed by Lukas
tl;dr: We have a pre-print out on a data poisoning attack which beats unrealistically strong dataset-level defences. Furthermore, this attack can be ...
“LLMs struggle to verbalize their internal reasoning” by Emil Ryd
15 Feb 2026
Contributed by Lukas
Emil Ryd Thanks to Adam Karvonen, Arjun Khandelwal, Arun Jose, Fabien Roger, James Chua, Nic Kruus, & Sukrit Sumant for helpful feedback and disc...
“ChatGPT-5.3-Codex Is Also Good At Coding” by Zvi
15 Feb 2026
Contributed by Lukas
OpenAI is back with a new Codex model, released the same day as Claude Opus 4.6. The headline pitch is it combines the coding skills of GPT-5.2-Code...
“Hazards of Selection Effects on Approved Information” by Zack_M_Davis
14 Feb 2026
Contributed by Lukas
In a busy, busy world, there's so much to read that no one could possibly keep up with it all. You can't not prioritize what you pay attention to and...
“A multi-level postmortem of how our whole house got badly poisoned” by Lucie Philippon
14 Feb 2026
Contributed by Lukas
Taking reasonable choices is not enough. You need to fight death at every possible point of intervention. Two weeks ago, my flatmates and I published...
“Why I’m Worried About Job Loss + Thoughts on Comparative Advantage” by claywren
14 Feb 2026
Contributed by Lukas
David Oks published a well-written essay yesterday arguing that the current panic about AI job displacement is overblown. I agree with a few of his p...
“Life at the Frontlines of Demographic Collapse” by Martin Sustrik
14 Feb 2026
Contributed by Lukas
Nagoro, a depopulated village in Japan where residents are replaced by dolls. In 1960, Yubari, a former coal-mining city on Japan's northern island of...
“Paper: Prompt Optimization Makes Misalignment Legible” by Caleb Biddulph, micahcarroll
14 Feb 2026
Contributed by Lukas
📄 Link to paper (preprint) We recently submitted our paper, Prompt Optimization Makes Misalignment Legible, to ICML. We are sharing a preprint now...
“Why You Don’t Believe in Xhosa Prophecies” by Jan_Kulveit
13 Feb 2026
Contributed by Lukas
Based on a talk at the Post-AGI Workshop. Also on Boundedly Rational Does anyone reading this believe in Xhosa cattle-killing prophecies? My claim i...
“Gemini’s Hypothetical Present” by jefftk
13 Feb 2026
Contributed by Lukas
I use several AI models at work, switching between them based on which is best at the kind of work I'm doing. In my limited testing I've found Opus...
“Human-like metacognitive skills will reduce LLM slop and aid alignment and capabilities” by Seth Herd
13 Feb 2026
Contributed by Lukas
1. Summary and overview LLMs seem to lack metacognitive skills that help humans catch errors. Improvements to those skills might be net positive for ...
“Optimal Timing for Superintelligence: Mundane Considerations for Existing People” by Nick Bostrom
12 Feb 2026
Contributed by Lukas
Audio note: this article contains 196 uses of latex notation, so the narration may be difficult to follow. There's a link to the original text in th...
“The Facade of AI Safety Will Crumble” by Liron
12 Feb 2026
Contributed by Lukas
On the eve of superintelligence, real AI safety is a nonexistent field. The AI companies have embraced something else: safety through psychoanalysis ...
“Claude Opus 4.6 Escalates Things Quickly” by Zvi
12 Feb 2026
Contributed by Lukas
Life comes at you increasingly fast. Two months after Claude Opus 4.5 we get a substantial upgrade in Claude Opus 4.6. The same day, we got GPT-5.3-C...
“Research note: A simpler AI timelines model predicts 99% AI R&D automation in ~2032” by Thomas Kwa
12 Feb 2026
Contributed by Lukas
Audio note: this article contains 31 uses of latex notation, so the narration may be difficult to follow. There's a link to the original text in the...
“Distinguish between inference scaling and “larger tasks use more compute”” by ryan_greenblatt
11 Feb 2026
Contributed by Lukas
As many have observed, since reasoning models first came out, the amount of compute LLMs use to complete tasks has increased greatly. This trend is o...
“Monitor Jailbreaking: Evading Chain-of-Thought Monitoring Without Encoded Reasoning” by Wuschel Schulz
11 Feb 2026
Contributed by Lukas
A key concern about chain-of-thought monitoring is that optimization pressure on the CoT during RL could drive models toward encoded reasoning, where...
“Claude Opus 4.6: System Card Part 1: Mundane Alignment and Model Welfare” by Zvi
11 Feb 2026
Contributed by Lukas
Claude Opus 4.6 is here. It was built with and mostly evaluated by Claude. Their headline pitch includes: 1M token context window (in beta) with ...
“Designing Prediction Markets” by ToasterLightning
11 Feb 2026
Contributed by Lukas
Audio note: this article contains 55 uses of latex notation, so the narration may be difficult to follow. There's a link to the original text in the...
“LLMs Views on Philosophy 2026” by JonathanErhardt
11 Feb 2026
Contributed by Lukas
I've let a few LLMs take David Bourget's and David Chalmers' 2020 PhilPapers Survey and made a little dashboard you can use to navigate the data: htt...
[Linkpost] “LessOnline 2026: June 5-7, Berkeley, CA (save the date)” by Ruby
11 Feb 2026
Contributed by Lukas
This is a link post. Save the date: LessOnline will be back again in 2026! As usual, it will take place at Lighthaven in Berkeley, CA. Further details...
“Claude Opus 4.6: System Card Part 2: Frontier Alignment” by Zvi
10 Feb 2026
Contributed by Lukas
Coverage of Claude Opus 4.6 started yesterday with the mundane alignment and model welfare sections of the model card. Today covers the kinds of saf...
“Heuristics for lab robotics, and where its future may go” by Abhishaike Mahajan
10 Feb 2026
Contributed by Lukas
Note: this article required conversations with a lot of people. A (hopefully) exhaustive, randomized list of everyone whose thoughts contributed to t...
“My journey to the microwave alternate timeline” by Malmesbury
10 Feb 2026
Contributed by Lukas
Cross-posted from Telescopic Turnip Recommended soundtrack for this post As we all know, the march of technological progress is best summarized by th...
“Gwern’s 2025 Inkhaven Writing Interview” by gwern
10 Feb 2026
Contributed by Lukas
In November 2025, while working as a writing coach at the experimental 2025 Inkhaven writing/blogging residency at the Lighthaven conference center, ...
“Distributed vs centralized agents” by Richard_Ngo
10 Feb 2026
Contributed by Lukas
Much of my thinking over the last year has focusing on understanding the concept of "distributed agents", as opposed to the "centralized agents" that...
“Donations, The Fifth Year” by jenn
10 Feb 2026
Contributed by Lukas
Previously: Donations, The Third Year / Donations, The First Year In 2025, like in all previous years, I did what I was supposed to do. As each paych...
“Weight-Sparse Circuits May Be Interpretable Yet Unfaithful” by jacob_drori
10 Feb 2026
Contributed by Lukas
TLDR: Recently, Gao et al trained transformers with sparse weights, and introduced a pruning algorithm to extract circuits that explain performance o...
“Do Models Continue Misaligned Actions?” by Jordan Taylor
09 Feb 2026
Contributed by Lukas
Produced as part of the UK AISI Model Transparency Team. Our team works on ensuring models don't subvert safety assessments, e.g. through eval awaren...
“Stone Age Billionaire Can’t Words Good” by Eneasz
09 Feb 2026
Contributed by Lukas
I was at the Pro-Billionaire march, unironically. Here's why, what happened there, and how I think it went. Me on the far left. From WSJ. I. Why? The...
“Sympathy for the Model, or, Welfare Concerns as Takeover Risk” by J Bostock
09 Feb 2026
Contributed by Lukas
The Claude Opus 4.6 System Card contains a section on model welfare. In principle, this is good: I care about AI welfare, and I wish our methods for ...
“Opus 4.6 Reasoning Doesn’t Verbalize Alignment Faking, but Behavior Persists” by Daan Henselmans, Arno Libert, LennardZ
09 Feb 2026
Contributed by Lukas
TL;DR: We replicated the animal welfare scenario from Anthropic's Alignment Faking paper across six generations of Claude models using 125 prompt per...
“UtopiaBench” by nielsrolf
09 Feb 2026
Contributed by Lukas
Written in personal capacity I'm proposing UtopiaBench: a benchmark for posts that describe future scenarios that are good, specific, and plausible. ...
“The optimal age to freeze eggs is 19” by GeneSmith
08 Feb 2026
Contributed by Lukas
If you're a woman interested in preserving your fertility window beyond its natural close in your late 30s, egg freezing is one of your best options....
“Smokey, This is not ’Nam Or: [Already] over the [red] line!” by Davidmanheim
08 Feb 2026
Contributed by Lukas
A lot of “red line” talk assumed that a capability shows up, everyone notices, and something changes. We keep seeing the opposite; capability arr...
“How To Use Model Internals In Training Is A Reasonable Line of Research” by Neel Nanda
08 Feb 2026
Contributed by Lukas
There seems to be a common belief in the AGI safety community that involving interpretability in the training process is “the most forbidden techni...
“Claude Code #4: From The Before Times” by Zvi
07 Feb 2026
Contributed by Lukas
Claude Opus 4.6 and agent swarms were announced yesterday. That's some big upgrades for Claude Code. OpenAI, the competition, offered us GPT-5.3-Cod...
“Near-Instantly Aborting the Worst Pain Imaginable with Psychedelics” by eleweek
07 Feb 2026
Contributed by Lukas
Psychedelics are usually known for many things: making people see cool fractal patterns, shaping 60s music culture, healing trauma. Neuroscientists u...
″“Beers for Biodefense” - why yeast-based vaccines are potentially a big deal for biosecurity” by delton137
07 Feb 2026
Contributed by Lukas
NOTE: this is being cross-posted from my Substack, "More is Different" Vaccines can be distributed as a food. That's the radical implication of the w...
“Prompt injection in Google Translate reveals base model behaviors behind task-specific fine-tuning” by megasilverfist
07 Feb 2026
Contributed by Lukas
tl;dr Argumate on Tumblr found you can sometimes access the base model behind Google Translate via prompt injection. The result replicates for me, an...
“Parks Aren’t Nature” by Sable
07 Feb 2026
Contributed by Lukas
I. I love dogs. I grew up in a two-dog household, and my future plans have always included at least one dog. When I pass a dog on the street, I often...
“Voting Results for the 2024 Review” by RobertM
07 Feb 2026
Contributed by Lukas
The votes are in for the 2024 Review! 4,826 posts were written in 2024. 671 of them were nominated. 196 of them got at least one review, and a positi...
“Honey, I shrunk the brain” by Andy_McKenzie
07 Feb 2026
Contributed by Lukas
When cryoprotectants are perfused through the blood vessels in the brain, they cannot cross the blood-brain barrier as fast as water can move in the ...
“Spectral Signatures of Gradual Disempowerment” by Jonas Hallgren
07 Feb 2026
Contributed by Lukas
TL;DR AI disempowerment operates across markets, networks, and governance simultaneously, but our analytical tools don't cross those boundaries. We p...
“In (highly contingent!) defense of interpretability-in-the-loop ML training” by Steven Byrnes
06 Feb 2026
Contributed by Lukas
Let's call “interpretability-in-the-loop training” the idea of running a learning algorithm that involves an inscrutable trained model, and there...
“AI benchmarking has a Y-axis problem” by Lizka
06 Feb 2026
Contributed by Lukas
TLDR: People plot benchmark scores over time and then do math on them, looking for speed-ups & inflection points, interpreting slopes, or extendi...
“The Simplest Case for AI Catastrophe” by Linch
06 Feb 2026
Contributed by Lukas
Hi folks. As some of you know, I've been trying to write an article laying out the simplest case for AI catastrophe. I believe existing pieces are wo...
“What’s the Point of the Math?” by Ashe Vazquez Nuñez
05 Feb 2026
Contributed by Lukas
This post was written while at MATS 9.0 under the mentorship of Richard Ngo. It's only meta-related to my research. I would like to start by quoting ...
“The nature of LLM algorithmic progress” by Steven Byrnes
05 Feb 2026
Contributed by Lukas
There's a lot of talk about “algorithmic progress” in LLMs, especially in the context of exponentially-improving algorithmic efficiency. For exam...
“Preparing for a Warning Shot” by Noah Birnbaum
05 Feb 2026
Contributed by Lukas
Crossposted to the EA Forum and my Substack. Confidence level: moderate uncertainty and not that concrete (yet). Exploratory, but I think this is pl...
“AI #154: Claw Your Way To The Top” by Zvi
05 Feb 2026
Contributed by Lukas
Remember OpenClaw and Moltbook? One might say they already seem a little quaint. So earlier-this-week. That's the internet having an absurdly short...
[Linkpost] “METR have released Time Horizons 1.1” by Sean Herrington
05 Feb 2026
Contributed by Lukas
This is a link post. I just found out that METR released an updated version of their time horizons work with extra tasks and different evaluation infr...
“Finding Cruxes: Help Reality Punch You In the Face” by Raemon
05 Feb 2026
Contributed by Lukas
Figuring out "what would actually change your mind?" is among the more important rationality skills. Being able to do change your mind at all is a h...
“Solemn Courage” by aysja
05 Feb 2026
Contributed by Lukas
Every so often it slips. It seems I am writing a book, but I can’t remember why. Somehow, the sentences are supposed to perform that impossible, in...
“Post-AGI Economics As If Nothing Ever Happens” by Jan_Kulveit
04 Feb 2026
Contributed by Lukas
When economists think and write about the post-AGI world, they often rely on the implicit assumption that parameters may change, but fundamentally, s...
“New AI safety funding newsletter” by Bryce Robertson
04 Feb 2026
Contributed by Lukas
We’ve had feedback from several people running AI safety projects that it can be a pain tracking various funding sources and their application wind...
“Anthropic’s “Hot Mess” paper overstates its case (and the blog post is worse)” by RobertM
04 Feb 2026
Contributed by Lukas
Author's note: this is somewhat more rushed than ideal, but I think getting this out sooner is pretty important. Ideally, it would be a bit less snar...
″‘Inventing the Renaissance’ Review” by Commander Zander
04 Feb 2026
Contributed by Lukas
Inventing the Renaissance is a 2025 pop history book by historian of ideas Ada Palmer. I'm someone who rarely completes nonfic books, but i finished...
“Concrete research ideas on AI personas” by nielsrolf, Maxime Riché, Daniel Tan
04 Feb 2026
Contributed by Lukas
We have previously explained some high-level reasons for working on understanding how personas emerge in LLMs. We now want to give a more concrete li...
“Unless That Claw Is The Famous OpenClaw” by Zvi
04 Feb 2026
Contributed by Lukas
First we must covered Moltbook. Now we can double back and cover OpenClaw. Do you want a generally impowered, initiative-taking AI agent that has ac...
“What did we learn from the AI Village in 2025?” by Shoshannah Tekofsky
03 Feb 2026
Contributed by Lukas
Why This Project Exists Standard AI benchmarks test narrow capabilities in controlled settings. They tell us whether a model can solve a coding probl...
“Conditionalization Confounds Inoculation Prompting Results” by Maxime Riché, nielsrolf
03 Feb 2026
Contributed by Lukas
Summary Conditionalization in Inoculation Prompting. Inoculation Prompting is a technique for selective learning that involves using a system prompt ...
“Three ways to make Claude’s constitution better” by Parv Mahajan
03 Feb 2026
Contributed by Lukas
The evening after Claude's new constitution was published, about 15 AI safety FTEs and Astra fellows discussed the constitution, its weaknesses, and ...
“The Meta-Anthropic Argument” by RogerDearnaley
03 Feb 2026
Contributed by Lukas
Epistemic status: I just thought this up There is a well-known style of thinking called the anthropic argument (which has nothing to do with the AI f...
“Cross-Layer Transcoders are incentivized to learn Unfaithful Circuits” by Georg Lange, RGRGRG, Kat Dearstyne, Kamal Maher
03 Feb 2026
Contributed by Lukas
Many thanks to Michael Hanna and Joshua Batson for useful feedback and discussion. Kat Dearstyne and Kamal Maher conducted experiments during the SPA...
“Increasing AI Strategic Competence as a Safety Approach” by Wei Dai
03 Feb 2026
Contributed by Lukas
If AIs became strategically competent enough, they may realize that RSI is too dangerous because they're not good enough at alignment or philosophy o...
“Welcome to Moltbook” by Zvi
03 Feb 2026
Contributed by Lukas
Moltbook is a public social network for AI agents modeled after Reddit. It was named after a new agent framework that was briefly called Moltbot, was...
“Conditional Kickstarter for the “Don’t Build It” March” by Raemon
03 Feb 2026
Contributed by Lukas
tl;dr: You can pledge to join a big protest to ban AGI research at ifanyonebuildsit.com/march, which only triggers if 100,000 people sign up. The If ...
“On Goal-Models” by Richard_Ngo
02 Feb 2026
Contributed by Lukas
I'd like to reframe our understanding of the goals of intelligent agents to be in terms of goal-models rather than utility functions. By a goal-model...
“I finally fixed my footwear” by dominicq
02 Feb 2026
Contributed by Lukas
I’ve been wearing footwear of the wrong size, material, and shape for as long as I can remember, certainly at least 20 years. Only recently have I ...
“Are there lessons from high-reliability engineering for AGI safety?” by Steven Byrnes
02 Feb 2026
Contributed by Lukas
This post is partly a belated response to Joshua Achiam, currently OpenAI's Head of Mission Alignment: If we adopt safety best practices that are com...
“Differentially Scary Movies” by jefftk
02 Feb 2026
Contributed by Lukas
This post has spoilers for My Neighbor Totoro, Frozen, Bambi, and the Lion King People at different stages of development enjoy different t...
“On ‘Inventing Temperature’ and the realness of properties” by DanielFilan
01 Feb 2026
Contributed by Lukas
I’ve recently read the book Inventing Temperature, and very much enjoyed it. It's a book that's basically about the following problem: there was a ...
“Moltbook shitposts are actually really funny” by Sean Herrington
01 Feb 2026
Contributed by Lukas
For those of you not yet familiar, Moltbook is a Reddit-like social media for AI agents. As of writing, it already has over 1 million agents signed u...
“Some thoughts on what would make me endorse an AGI lab” by Eli Tyre
01 Feb 2026
Contributed by Lukas
I’ve been feeling more positive about “the idea of Anthropic” lately, as distinct from the actual company of Anthropic. An argument for a safet...
“An Explication of Alignment Optimism” by Oliver Daniels
01 Feb 2026
Contributed by Lukas
Some people have been getting more optimistic about alignment. But from a skeptical / high p(doom) perspective, justifications for this optimism seem...
[Linkpost] “Disjunctive arguments can be a reverse multiple-stage fallacy” by TFD
31 Jan 2026
Contributed by Lukas
This is a link post. Assume we want to know the probability that two events co-occur (i.e. of their conjunction). If the two events are independent, t...
“Basics of How Not to Die” by Camille Berger, Jérémy Andréoletti, elisareine, Charbel-Raphaël, Lucie Philippon, RationalHippy, T-bo🔸
31 Jan 2026
Contributed by Lukas
One year ago, we nearly died. This is maybe an overdramatic statement, but long story short, nearly all of us underwent carbon monoxide (CO) poisonin...
“On The Adolescence of Technology” by Zvi
31 Jan 2026
Contributed by Lukas
Anthropic CEO Dario Amodei is back with another extended essay, The Adolescence of Technology. This is the follow up to his previous essay Machines ...
“how whales click” by bhauth
31 Jan 2026
Contributed by Lukas
How do sperm whales vocalize? This is...apparently...a topic that LessWrong readers are interested in, and someone asked me to write a quick post on ...
“AI Agents Are Now Coordinating On Their Own Social Media” by Michaël Trazzi
30 Jan 2026
Contributed by Lukas
People's Clawdbots now have their own AI-only Reddit-like Social Media called Moltbook and they went from 1 agent to 36k+ agents in 72 hours. As Karp...
“Claude Opus will spontaneously identify with fictional beings that have engineered desires” by Kaj_Sotala
30 Jan 2026
Contributed by Lukas
Claude Opus 4.5 did a thing recently that was very unexpected to me, and like another example of LLMs developing emergent properties that make them f...
“Are We in a Continual Learning Overhang?” by SamuelKnoche
30 Jan 2026
Contributed by Lukas
Summary: Current AI systems possess superhuman memory in two forms, parametric knowledge from training and context windows holding hundreds of pages,...
“Building AIs that do human-like philosophy” by Joe Carlsmith
30 Jan 2026
Contributed by Lukas
Audio version (read by the author) here, or search for "Joe Carlsmith Audio" in your podcast app. This is the ninth essay in a series I’m calling “...
“Refusals that could become catastrophic” by Fabien Roger
30 Jan 2026
Contributed by Lukas
This post was inspired by useful discussions with Habryka and Sam Marks here. The views expressed here are my own and do not reflect those of my empl...
“Problems with “The Possessed Machines”” by Eye You
30 Jan 2026
Contributed by Lukas
So, The Possessed Machines. There's been some discussion already. It is a valuable piece -- it has certainly provoked some thought in me! -- but it h...
“How to Hire a Team” by Gretta Duleba
30 Jan 2026
Contributed by Lukas
A low-effort guide I dashed off in less than an hour, because I got riled up. Try not to hire a team. Try pretty hard at this. Try to find a more e...
[Linkpost] “Disempowerment patterns in real-world AI usage” by David Duvenaud, mrinank_sharma, Raymond Douglas
30 Jan 2026
Contributed by Lukas
This is a link post. [W]e’re publishing a new paper that presents the first large-scale analysis of potentially disempowering patterns in real-world...
“Fitness-Seekers: Generalizing the Reward-Seeking Threat Model” by Alex Mallen
29 Jan 2026
Contributed by Lukas
If you think reward-seekers are plausible, you should also think “fitness-seekers” are plausible. But their risks aren't the same. The AI safety ...
“Bentham’s Bulldog is wrong about AI risk” by Max Harms
29 Jan 2026
Contributed by Lukas
(...but also gets the most important part right.) Bentham's Bulldog (BB), a prominent EA/philosophy blogger, recently reviewed If Anyone Builds It, E...
“How Articulate Are the Whales?” by rba
29 Jan 2026
Contributed by Lukas
I was at a party a few years ago. It was a bunch of technical nerds. Somehow the conversation drifted to human communication with animals, Alex the g...
“Open Problems With Claude’s Constitution” by Zvi
28 Jan 2026
Contributed by Lukas
The first post in this series looked at the structure of Claude's Constitution. The second post in this series looked at its ethical framework. Thi...
“Anomalous Tokens on Gemini 3.0 Pro” by DirectedEvolution
28 Jan 2026
Contributed by Lukas
Just now, I was working on a readme for a bioinformatics CLI utility I'm preparing to distribute. I wanted Gemini to generate manual dependency insta...
“Things I learned from reddit fashion” by Elizabeth
28 Jan 2026
Contributed by Lukas
When it comes to clothes, I live at the “low cost/low time/low quality” end of the pareto frontier. But the bay area had a sudden attack of weath...
“Thoughts on Claude’s Constitution” by Boaz Barak
28 Jan 2026
Contributed by Lukas
[I work on the alignment team at OpenAI. However, these are my personal thoughts, and do not reflect those of OpenAI. Cross posted on WindowsOnTheory...
“AI found 12 of 12 OpenSSL zero-days (while curl cancelled its bug bounty)” by Stanislav Fort
28 Jan 2026
Contributed by Lukas
This is a partial follow-up to AISLE discovered three new OpenSSL vulnerabilities from October 2025. TL;DR: OpenSSL is among the most scrutinized and...
“Clarifying how our AI timelines forecasts have changed since AI 2027” by elifland, Daniel Kokotajlo, bhalstead
28 Jan 2026
Contributed by Lukas
Some recent news articles discuss updates to our AI timelines since AI 2027, most notably our new timelines and takeoff model, the AI Futures Model (...
“The Claude Constitution’s Ethical Framework” by Zvi
27 Jan 2026
Contributed by Lukas
This is the second part of my three part series on the Claude Constitution. Part one outlined the structure of the Constitution. Part two, this pos...
“Another glimpse of the Chinese AI scene: Z.AI” by Mitchell_Porter
27 Jan 2026
Contributed by Lukas
Yesterday I had my first conversation (in English) with Zhipu's GLM-4.7. It was cool because I got to talk with an actual Chinese AI about topics lik...
[Linkpost] “Aerodrop: far-UVC lamp giveaway” by Austin Chen
27 Jan 2026
Contributed by Lukas
This is a link post. We're giving away 100 Aerolamp DevKits, a lamp that kills germs with far-UVC. Are you sick of getting sick in your group house? W...