To understand AI, you should use it. Here’s how to get started. By Peter Hartree · Published April 4th, 2025 · On this page: To truly understand what AI can do — and what is coming soon — you should make regular use of the latest AI services. This article will help you learn how to do this. Table of…

It’s been a crazy few weeks here at the AI Futures Project. Almost a million people visited our webpage; 166,000 watched our Dwarkesh interview. We were invited on something like a million podcasts. Team members gave talks at Harvard, the Federation of American Scientists, and OpenAI. People even -…

When I was a really small kid, one of my favorite activities was to try and dam up the creek in my backyard. I would carefully move rocks into high walls, pile up leaves, or try patching the holes with sand. The goal was just to see how high I could get the lake, knowing that if I plugged every…

Podcast version here (read by the author), or search for "Joe Carlsmith Audio" on your podcast app. This is the fourth essay in a series that I’m calling “How do we solve the alignment problem?”. I’m hoping that the individual essays can be read fairly well on their own, but see this introduction…

My goal as an AI safety researcher is to put myself out of a job. I don’t worry too much about how planet sized brains will shape galaxies in 100 years. That’s something for AI systems to figure out. Instead, I worry about safely replacing human researchers with AI agents, at which point human…

Lecture by Joe Carlsmith on his essay series "Otherness and control in the age of AGI." Stanford University, CS 362: Research in AI Alignment, 1 October 2024. Transcript here: https://joecarlsmith.com/2024/10/08/video-and-transcript-of-presentation-on-otherness-and-control-in-the-age-of-agi Essay…

LESSWRONG LW Laziness death spirals 19th Sep 2024 10 min read 280 New Comment 40 comments, sorted by Click to highlight new comments since: cubefox10mo2918 Other options: Writing a to-do list. If the tasks on the list are too daunting, break them apart into smaller tasks. Changing location to an…

LESSWRONG LW New Comment 11 comments, sorted by Click to highlight new comments since: Tao Lin26dΩ684 This overestimates the impact of large models on external safety research. My impression is that the AI safety community has barely used deepseek r1 and v3 open source weights at all. I checked…

John Wentworth recently posted a discussion-inducing piece about how the willingness to be vulnerable is core to romantic relationships. This is not obvious to me as the core thing. I think romantic (and other) relationships are arguably about intimacy. Perhaps intimacy always requires…

Epistemic status: Vague and handwavy. Nuance is often missing. Some of the claims depend on implicit definitions that may be reasonable to disagree with. But overall I think it's directionally true.

Sometimes people say that it's much less valuable to do AI safety research today than it will be in the future, because the current models are very different—in particular, much less capable—than the models that we're actually worried about. I think that argument is mostly right, but it misses a…

One risk of “vibe-coding” a piece of software with an LLM is that it gets you 90% of the way there, but then you’re stuck—the last 10% of bug fixes, performance improvements, or additional features is really hard to figure out because the AI has written messy, verbose code that both of you struggle…

isolated demand for rigor

dynomight.net/blood In a recent post about trading stuff for money, I mentioned: Europe had a [blood plasma] shortage of around 38%, which it met by importing plasma from paid donors in the United States, where blood products account for 2% of all exports by value. The internet’s reaction was: “TWO…

For months, I had the feeling: something is wrong. Some core part of myself had gone missing. I had words and ideas cached, which pointed back to the missing part. There was the story of Benjamin Jesty, a dairy farmer who vaccinated his family against smallpox in 1774 - 20 years before the…

[previously in series: 1, 2, 3, 4, 5] When that April with his sunlight fierce The rainy winter of the coast doth pierce And filleth every spirit with such hale As horniness engenders in the male Then folk go out in crop tops and in shorts Their bodies firm from exercise and sports And men gaze at…

[previously in series: 1, 2, 3, 4] It has been three weeks since Sam Altman was fired, but the conversation won’t move on. “What did Ilya see?” asks your Uber driver, on the way to the airport. “What

[previously in series: 1, 2, 3] You spent the evening agonizing over which Bay Area House Party to attend. The YIMBY parties are always too crowded. VC parties were a low-interest-rate phenomenon. You

[Previously: Every Bay Area House Party, Another Bay Area House Party] People talk about “fuck-you money”, the amount you’d have to make to never work again. You dream of fuck-you social success, wher

[Previously: Every Bay Area House Party] Blaise Pascal said all human evil comes from inability to sit alone in a room. Your better nature - your rational soul - tells you that nothing good has ever c

Some of the more unhinged writing on superintelligence pictures AI doing things that seem like magic. Crossing air gaps to escape its data center. Building nanomachines from simple components. Plowing through physical bottlenecks to revolutionize the economy in months. More sober thinkers point out…

dynomight.net/money Examples are good. Let’s start with some examples: We all need kidneys, or at least one kidney. Donating a kidney sucks, but having zero working kidneys really sucks. Paying people for kidneys would increase the number available, but it seems gross to pay people for part of their…

We’re not the only people who get confused by OpenAI model names. There’s actually a simple explanation: most of the models are modifications of each other in a giant dependency graph that OpenAI is keeping secret. In this post we’ll dive into explaining our best guess on how that graph looks.1…

For a lay audience, but I've seen a surprising number of knowledgeable people fretting over depressed-seeming comics from current systems. Either they're missing something or I am. Perhaps you’ve seen images like this self-portrait from ChatGPT, when asked to make a comic about its own experience.…

Our universe is probably a computer simulation created by a paperclip maximizer to map the spectrum of rival resource‑grabbers it may encounter while expanding through the cosmos. The purpose of this simulation is to see what kind of ASI (artificial superintelligence) we humans end up creating. The…

Comp sci in 2017: Student: I get the feeling the compiler is just ignoring all my comments. Teaching assistant: You have failed to understand not just compilers but the concept of computation itself. Comp sci in 2027: Student: I get the feeling the compiler is just ignoring all my comments. TA:…

With an Elo rating of 2706, OpenAI’s o3 model is better than 99.7% of competitive programmers on CodeForces (based on this rating distribution data from 2024). Humans with this performance would excel at any top tech company, whereas o3 is not ready to automate even the median big-tech engineer’s…

With an Elo rating of 2706, OpenAI’s o3 model is better than 99.7% of competitive programmers on CodeForces (based on this rating distribution data from 2024). Humans with this performance would excel at any top tech company, whereas o3 is not ready to automate even the median big-tech engineer’s…

Britain today celebrates the crowning of a new king, but the coronation will be watched and celebrated by millions across the Commonwealth. To an extent that is often not appreciated abroad, the Queen – who was Defender of the Faith – was revered by her subjects of all faiths. In our often sectarian…

If you don’t have time to read this post, these five graphs give most of the argument. Each includes both the energy/water cost of using ChatGPT in the moment and the amortized cost of training GPT-4: Source for original graph. Each bar here represents 1 year of the activity, so the live car-free…

Image: Cursor In a viral video posted on X, an 8-year-old girl happily “codes” away. Over the course of 45 minutes, she has Cursor, an AI coding assistant, build a chatbot that replies in the style of Harry Potter. Using only simple text-based prompts, she asks Cursor to create a website, add a…

The Eve of the Deluge. (John Martin) (Part 1 here) The AGI frog is getting boiled As mentioned in part 1, a brief market correction happened in late 2026. In 2027, OpenAI releases o7 to try to shore up excitement and new investments. It’s much more reliable than o6 and can now do a lot of office…

The Great Day of His Wrath. (John Martin) (Part 1 here, part 2 here) The end of white-collar work and the new job scene By the late 2020s, office jobs in developed countries are now basically about overseeing and providing direction to AI systems, and the last part of that is mostly on-paper rather…

Below is part 1 of an extended scenario describing how the future might go if current trends in AI continue. The scenario is deliberately extremely specific: it’s definite rather than indefinite, and makes concrete guesses instead of settling for banal generalities or abstract descriptions of…

Most people think of AI as a pattern-matching chatbot – good at writing emails, terrible at real thinking. They've missed something huge. In 2024, while many declared AI was reaching a plateau, it was actually entering a new paradigm: learning to reason using reinforcement learning. This approach…

This post should not be taken as a polished recommendation to AI companies and instead should be treated as an informal summary of a worldview. The content is inspired by conversations with a large number of people, so I cannot take credit for any of these ideas. For a summary of this post, see the…

The AI Control Agenda, in its own words: … we argue that AI labs should ensure that powerful AIs are controlled. That is, labs should make sure that the safety measures they apply to their powerful models prevent unacceptably bad outcomes, even if the AIs are misaligned and intentionally try to…

As noted further above, most the advice above is selected from the resources linked below (the list is in no particular order):