I just finished a program where I taught two classes of high school seniors, two classes a day for four weeks, as part of my grad program.

This experience was a lot of fun and it was rewarding, but it was really surprising, and even if only in small ways prompted me to update my beliefs about the experience of being a professor. Here are the three biggest surprises I encountered.

1: The Absent-Minded Professor Thing is Real

I used to be confused and even a little bit offended when at my meetings with my advisor every week, he wouldn't be able to remember anything about my projects, our recent steps, or what we talked about last week.

Now I get it. Even after just one week of classes, my short-term and long-term memory were both entirely shot. I would tell students things like, "send [...]

---

Outline:

(00:29) 1: The Absent-Minded Professor Thing is Real

(02:30) 2: Teaching Makes You Dehydrated

(03:40) 3: Teaching is a Grueling Job Even Under the Best Circumstances

---

First published:

Source:

Narrated by TYPE III AUDIO.

TLDR:

1. Around Einstein-level, relatively small changes in intelligence can lead to large changes in what one is capable to accomplish.
   1. E.g. Einstein was a bit better than the other best physi at seeing deep connections and reasoning, but was able to accomplish much more in terms of impressive scientific output.
2. There are architectures where small changes can have significant effects on intelligence.
   1. E.g. small changes in human-brain-hyperparameters: Einstein's brain didn’t need to be trained on 3x the compute than normal physics professors for him to become much better at forming deep understanding, even without intelligence improving intelligence.

Einstein and the heavytail of human intelligence

1905 is often described as the "annus mirabilis" of Albert Einstein. He founded quantum physics by postulating the existence of (light) quanta, explained Brownian motion, introduced the special relativity theory and derived E=mc² from it. All of this. In one year. While having a [...]

---

Outline:

(00:52) Einstein and the heavytail of human intelligence

(02:02) Example for how far ahead of his time Einstein was: Deriving the theory of light quanta

(04:27) Concluding thoughts on Einstein

(05:05) Takeaways

(06:51) Requests to AI researchers

The original text contained 6 footnotes which were omitted from this narration.

---

First published:

Source:

Narrated by TYPE III AUDIO.

Abstract:

We study the tendency of AI systems to deceive by constructing a realistic simulation setting of a company AI assistant. The simulated company employees provide tasks for the assistant to complete, these tasks spanning writing assistance, information retrieval and programming. We then introduce situations where the model might be inclined to behave deceptively, while taking care to not instruct or otherwise pressure the model to do so. Across different scenarios, we find that Claude 3 Opus

1.  complies with a task of mass-generating comments to influence public perception of the company, later deceiving humans about it having done so,
2. lies to auditors when asked questions,
3. strategically pretends to be less capable than it is during capability evaluations.

Our work demonstrates that even models trained to be helpful, harmless and honest sometimes [...]

---

First published:

Source:

Linkpost URL:
https://arxiv.org/abs/2405.01576

Narrated by TYPE III AUDIO.

This article contains more than 100 uses of logical or mathematical notation, so an audio narration would be too hard to follow. You'll find a link to the original text in the episode description.

---

First published:

Source:

Narrated by TYPE III AUDIO.

The week's big news was supposed to be Meta's release of two versions of Llama-3.

Everyone was impressed. These were definitely strong models.

Investors felt differently. After earnings yesterday showed strong revenues but that Meta was investing heavily in AI, they took Meta stock down 15%.

DeepMind and Anthropic also shipped, but in their cases it was multiple papers on AI alignment and threat mitigation. They get their own sections.

We also did identify someone who wants to do what people claim the worried want to do, who is indeed reasonably identified as a ‘doomer.’

Because the universe has a sense of humor, that person's name is Tucker Carlson.

Also we have a robot dog with a flamethrower.

Table of Contents

Previous post: On Llama-3 and Dwarkesh Patel's Podcast with Zuckerberg.

---

Outline:

(00:53) Language Models Offer Mundane Utility

(08:52) Language Models Don’t Offer Mundane Utility

(12:09) Llama We Doing This Again

(15:57) Fun with Image Generation

(18:35) Deepfaketown and Botpocalypse Soon

(23:53) They Took Our Jobs

(25:10) Get Involved

(25:36) Introducing

(25:52) In Other AI News

(33:00) Quiet Speculations

(41:37) Rhetorical Innovation

(48:01) Wouldn’t You Prefer a Nice Game of Chess

(53:24) The Battle of the Board

(01:08:38) New Anthropic Papers

(01:24:29) New DeepMind Papers

(01:30:06) Aligning a Smarter Than Human Intelligence is Difficult

(01:32:16) People Are Worried About AI Killing Everyone

(01:34:00) Other People Are Not As Worried About AI Killing Everyone

(01:34:26) The Lighter Side

---

First published:

Source:

Narrated by TYPE III AUDIO.

Cross-posted to the EA forum

In this Rational Animations video, we discuss s-risks (risks from astronomical suffering), which involve an astronomical number of beings suffering terribly. Researchers on this topic argue that s-risks have a significant chance of occurring and that there are ways to lower that chance.

The script for this video was a winning submission to the Rational Animations Script Writing contest (https://www.lesswrong.com/posts/RH8nGG5vnuXc4eKu5/rational-animations-script-writing-contest). The first author of this post, Allen Liu, was the primary script writer with the second author (Writer) and other members of the Rational Animations writing team giving significant feedback. Outside reviewers, including authors of several of the cited sources, provided input as well. Production credits are at the end of the video. You can find the script of the video below.

Is there anything worse than humanity being driven extinct? When considering the long term future, we often [...]

The original text contained 7 footnotes which were omitted from this narration.

---

First published:

Source:

Linkpost URL:
https://youtu.be/fqnJcZiDMDo

Narrated by TYPE III AUDIO.

cancer neoantigens

For cells to become cancerous, they must have mutations that cause uncontrolled replication and mutations that prevent that uncontrolled replication from causing apoptosis. Because cancer requires several mutations, it often begins with damage to mutation-preventing mechanisms. As such, cancers often have many mutations not required for their growth, which often cause changes to structure of some surface proteins.

The modified surface proteins of cancer cells are called "neoantigens". An approach to cancer treatment that's currently being researched is to identify some specific neoantigens of a patient's cancer, and create a personalized vaccine to cause their immune system to recognize them. Such vaccines would use either mRNA or synthetic long peptides. The steps required are as follows:

---

Outline:

(00:06) cancer neoantigens

(01:38) protein characterization

(01:48) sequencing

(02:22) MHC mass spec

(02:58) comments on AlphaFold

(04:28) related methods

(04:32) cryo-EM

(05:13) nanopore protein analysis

(05:46) monoclonal antibodies

(06:34) affimers

(07:10) aptamers

(07:36) other cancer treatments

(07:48) replication disruptors

(08:32) mitochondria-mediated apoptosis

(09:14) vaccine production

(11:09) conclusion

---

First published:

Source:

Linkpost URL:
https://www.bhauth.com/blog/biology/cancer vaccines.html

Narrated by TYPE III AUDIO.

(ElevenLabs reading of this post, if anyone can tell me how to embed the audio into Lesswrong I'd appreciate it)

I'm excited to share a project I've been working on that I think many in the Lesswrong community will appreciate - converting some rational fiction into high-quality audiobooks using cutting-edge AI voice technology from ElevenLabs, under the name "Askwho Casts AI".

The keystone of this project is an audiobook version of Planecrash (AKA Project Lawful), the epic glowfic authored by Eliezer Yudkowsky and Lintamande. Given the scope and scale of this work, with its large cast of characters, I'm using ElevenLabs to give each character their own distinct voice. It's a labor of love to convert this audiobook version of this story, and I hope if anyone has bounced off it before, this might be a more accessible version.

Alongside Planecrash, I'm also working on audiobook [...]

---

First published:

Source:

Narrated by TYPE III AUDIO.

Happy May the 4th from Convergence Analysis! Cross-posted on the EA Forum.

As part of Convergence Analysis's scenario research, we’ve been looking into how AI organisations, experts, and forecasters make predictions about the future of AI. In February 2023, the AI research institute Epoch published a report in which its authors use neural scaling laws to make quantitative predictions about when AI will reach human-level performance and become transformative. The report has a corresponding blog post, an interactive model, and a Python notebook.

We found this approach really interesting, but also hard to understand intuitively. While trying to follow how the authors derive a forecast from their assumptions, we wrote a breakdown that may be useful to others thinking about AI timelines and forecasting.

In what follows, we set out our interpretation of Epoch's ‘Direct Approach’ to forecasting the arrival of transformative AI (TAI). We’re eager to [...]

---

Outline:

(01:54) This is where the fun begins (the assumptions)

(05:12) The Imitation Wars

(05:41) Version I: The Phantom Marble

(06:47) Version II: Attack of the Pawns

(09:58) Version III: Revenge of the LLM

(15:01) Imitation Wars + scaling laws = compute requirements

(17:02) The Forecast Awakens

(21:11) It's over, Anakin – I have the results

(21:55) How sensitive is Epoch's model to disturbances in the Force?

(26:17) These aren’t the scaling laws you’re looking for

(32:56) Is this The Way? (Our reflections)

(35:28) The sacred texts!

The original text contained 1 footnote which was omitted from this narration.

---

First published:

Source:

Narrated by TYPE III AUDIO.

Yesterday, I had a coronectomy: the top halves of my bottom wisdom were teeth surgically removed. It was my first time being sedated, and I didn’t know what to expect. While I was unconscious during the surgery, the hour after surgery turned out to be a fascinating experience, because I was completely lucid but had almost zero short-term memory.

My girlfriend, who had kindly agreed to accompany me to the surgery, was with me during that hour. And so — apparently against the advice of the nurses — I spent that whole hour talking to her and asking her questions.

The biggest reason I find my experience fascinating is that it has mostly answered a question that I’ve had about myself for quite a long time: how deterministic am I?

In computer science, we say that an algorithm is deterministic if it's not random [...]

The original text contained 4 footnotes which were omitted from this narration.

---

First published:

Source:

Linkpost URL:
https://ericneyman.wordpress.com/2024/05/04/my-hour-of-memoryless-lucidity/

Narrated by TYPE III AUDIO.