Wiredframe Radar logo

Category intelligence

Research Briefing — December 31, 2025

14 current items analyzed and ranked.

Executive synthesis

Research Summary

A light batch for core AI research, with the strongest content in AI alignment education and biotechnology methodology.

Remaining items cover CFAR organizational updates, rationality workshop curricula, education policy proposals for gifted children, and science history writing. No traditional ML papers or benchmark results in this batch.

Key Themes

AI Safety & Alignment · 2Biotechnology · 2Programming & Human-AI Collaboration · 1CFAR & Rationality Community · 5Personal Finance & Philanthropy · 3

Primary evidence

Top Ranked Signals

Research LessWrong Dec 30

[Advanced Intro to AI Alignment] 1. Goal-Directed Reasoning and Why It Matters

By Towards_Keeperhood

52 score
AI Analysis
An educational introduction to AI alignment examining goal-directed reasoning through the 'thinking loop' framework (search, predict, evaluate, iterate). Connects these concepts to model-based reinforcement learning as a lens for understanding alignment challenges.
1.1 Summary and Table of ContentsWhy would an AI "want" anything? This post answers that question by examining a key part of the structure of intelligent cognition.When you solve a novel problem, your mind searches for plans, predicts their outcomes, evaluates whether those outcomes achieve what you want, and iterates. I call this the "thinking loop". We will build some intuition for why any AI capable of solving difficult real-world problems will need something structurally similar.This framewo
AI AlignmentReinforcement LearningAI SafetyGoal-Directed AI
Research LessWrong Dec 30

Chromosome identification methods

By TsviBT

45 score
AI Analysis
Technical proposal for chromosome identification methods to enable 'chromosome selection' - assembling genomes from chromosomes of multiple source cells. Proposes 'complementary identification' paradigm: destructively identify all chromosomes except one to infer the preserved one's identity.
PDF version. berkeleygenomics.org. x.com. bluesky. This is a linkpost for "Chromosome identification methods"; a few of the initial sections are reproduced here. Abstract Chromosome selection is a hypothetical technology that assembles the genome of a new living cell out of whole chromosomes taken from multiple source cells. To do chromosome selection, you need a method for chromosome identification—distinguishing between chromosomes by number, and ideally also by allele content. This article in
BiotechnologyGenomicsReproductive Technology
Research LessWrong Dec 30

Mechanize Work's essay on Unfalsifiable Doom

By StanislavKrym

42 score
AI Analysis
Commentary and critique of Mechanize Work's essay responding to Yudkowsky & Soares' book 'If Anyone Builds It, Everyone Dies', arguing that AI doom arguments are unfalsifiable and rely on vague theoretical arguments rather than concrete evidence. The post embeds the original text with inline commentary on potential errors.
Like Daniel Kokotajlo's coverage of Vitalik's response to AI-2027, I've copied the author's text. However, I would like to comment upon potential errors right in the text, since it would be clearer.Our critics tell us that our work will destroy the world.We want to engage with these critics, but there is no standard argument to respond to, no single text that unifies the AI safety community. Nonetheless, while this community lacks a central unifying argument, it does have a central figure:
AI SafetyAI Doom ArgumentsRationalist Discourse
Research LessWrong Dec 30

Many can write faster asm than the compiler, yet don't. Why?

By faul_sname

35 score
AI Analysis
Argues that while humans can often write faster assembly than compilers for specific cases, the real barrier is the difficulty of mixing hand-written and machine-generated assembly in practical projects. Draws implicit parallels to AI coding assistants.
There's a take I've seen going around, which goes approximately like this:It used to be the case that you had to write assembly to make computers do things, but then compilers came along. Now we have optimizing compilers, and those optimizing compilers can write assembly better than pretty much any human. Because of that, basically nobody writes assembly anymore. The same is about to be true of regular programming.I 85% agree with this take.However, I think there's one important inaccuracy: even
ProgrammingHuman-AI CollaborationSoftware Engineering
Research LessWrong Dec 30

Exceptionally Gifted Children

By John Boyle

25 score
AI Analysis
Talk and article about exceptionally gifted children (IQ 160+), proposing that schools allow subject-by-subject grade placement based on ability. Notes relevance to rationalist community given high average reader IQ.
I gave a talk on exceptionally gifted children at the Reproductive Frontiers Summit at Lighthaven this June.  I believe the subject matter is highly relevant to the experience of many rationalists (e.g. one of Scott's surveys has put the average IQ of his readers at 137, and although that's not as extreme as 160+, I think many of the observations generalize to the merely highly gifted).  The talk is on YouTube: I also adapted the talk into an article for the Center for Educational
EducationGifted ChildrenPolicy
Research LessWrong Dec 30

The origin of rot

By Abhishaike Mahajan

22 score
AI Analysis
A history-focused essay exploring the scientific understanding of decomposition and the thanatomicrobiome (microbial communities after death), using Saramago's novel 'Death with Interruptions' as a framing device.
Note: I spent my holidays writing a bunch of biology-adjacent, nontechnical pieces. I’ll intermittently mix them between whatever technical thing I send out, much like how a farmer may mix sawdust into feed, or a compounding pharmacist, butter into bathtub-created semaglutide. This one is about history!The book ‘Death with Interruptions’ is a 2005 speculative fiction novel written by Portuguese author José Saramago. It is about how, mysteriously, on January 1st of an unnamed year in an unnamed c
BiologyScience HistoryMicrobiology
Research LessWrong Dec 30

CFAR’s todo list re: our workshops

By AnnaSalamon

20 score
AI Analysis
CFAR organizational update listing known issues with their rationality workshops, including concerns that techniques may help with quick tasks but not sustained health behaviors like exercise or textbook learning.
(This post is part of a sequence of year-end efforts to invite real conversation about CFAR; you’ll find more about our workshops, as well as our fundraiser, at What’s going on at CFAR? Updates and Fundraiser and at More details on CFAR's new workshops)In part of that post, we discuss the main thing that bothered me about our past workshop and why I think it is probably fixed now (though we’re still keeping an eye out). Here, I list the biggest remaining known troubles with our workshops and our
CFARRationality TrainingCommunity Organizations
Research LessWrong Dec 30

What’s going on at CFAR? (Updates and Fundraiser)

By AnnaSalamon

20 score
AI Analysis
Comprehensive CFAR update and fundraiser post aiming for transparent community engagement, describing organizational approach and inviting feedback. Discusses how CFAR wants to be 'situated within a community that knows us.'
This post is the main part of a sequence of year-end efforts to invite real conversation about CFAR, published to coincide with our fundraiser. Introduction / What’s up with this postMy main aim with this post is to have a real conversation about aCFAR[1] that helps us be situated within a community that (after this conversation) knows us. My idea for how to do this is to show you guys a bunch of pieces of how we’re approaching things, in enough detail to let you kibitz.[2]My secondary
CFARCommunity OrganizationsFundraising
Research LessWrong Dec 30

The 7 Types Of Advice (And 3 Common Failure Modes)

By Linch

18 score
AI Analysis
A taxonomy of seven categories of advice (Master Key, shortcuts, guardrails, etc.) and three common failure modes when giving or receiving advice. Aimed at helping people identify what type of advice they need and avoid mismatches.
Reposting my Inkhaven post on ontology of advice here. Are you interested in learning a new field, whether it’s programming, writing, or how to win Paper Mario games? Have you searched for lots of advice and couldn’t choose which advice to follow? Worse, have you tried to follow other people’s Wise Sounding Advice and ended up worse than where you started?Alternatively, have you tried to give useful advice distilling your hard-earning learnings and only to realize it fell on deaf ears? Or p
RationalitySelf-ImprovementCommunication
Research LessWrong Dec 30

More details on CFAR’s new workshops

By AnnaSalamon

18 score
AI Analysis
Detailed description of CFAR's updated workshop format and new curriculum thread on 'who-ness' (ego, pride, authorship). Workshop maintains 4.5-day immersive format with ~25 participants.
(This post is part of a sequence of year-end efforts to invite real conversation about CFAR; you’ll find more about our workshops, as well as our fundraiser, at What’s going on at CFAR? Updates and Fundraiser.)If you’d like to know more about CFAR’s current workshops (either because you’re thinking of attending / sending a friend, or because you’re just interested), this post is for you. Our focus in this post is on the new parts of our content. Kibitzing on content is welcome and appreciated re
CFARRationality TrainingPersonal Development
Research LessWrong Dec 30

Don't Sell Stock to Donate

By jefftk

15 score
AI Analysis
Tax optimization advice explaining that donating appreciated stock directly to 501(c) organizations avoids capital gains tax, potentially increasing effective donation value by ~10%. Provides specific examples for different income levels and locations.
When you sell stock [1] you pay capital gains tax, but there's no tax if you donate the stock directly. Under a bunch of assumptions, someone donating $10k could likely increase their donations by ~$1k by donating stock. This applies to all 501(c) organizations, such as regular 501(c)3 non-profits, but also 501(c)4s such as advocacy groups. In the US, when something becomes more valuable and you sell it you need to pay tax proportional to the gains. [2] This gets complicated based on how much ot
Personal FinanceEffective AltruismPhilanthropy
Research LessWrong Dec 30

Dating Roundup #9: Signals and Selection

By Zvi

15 score
AI Analysis
Analysis of dating signals and selection effects, particularly examining 'the ick' as a Bayesian likelihood ratio update - small behaviors serving as evidence of deeper traits. Draws on research about disgust sensitivity and narcissism.
Ultimately, it comes down to one question. Are you in? For you, and for them. You’re Single Because They Got The Ick The Ick, the ultimate red flag, makes perfect sense and is all about likelihood ratios. Koenfucius: The ‘ick’ is a colloquial term for a feeling of disgust triggered by a specific—typically trivial—behaviour from a romantic partner, often leading to the relationship’s demise. New research explores why some are more prone to getting it than others. Robin Hanson: “Women also experie
Social DynamicsSignalingRationality