AI Wiki — Knowledge Base

from Write & Learn (thx @MilksandMatcha @philipkiely @swyx), my most important takeaway, 𝐛𝐮𝐢𝐥𝐝 𝐚 𝐝𝐞𝐧𝐬𝐞 𝐫𝐞𝐰𝐚𝐫𝐝 𝐟𝐨𝐫 𝐲𝐨𝐮…

from Write & Learn (thx @MilksandMatcha @philipkiely @swyx), my most important takeaway, 𝐛𝐮𝐢𝐥𝐝 𝐚 𝐝𝐞𝐧𝐬𝐞 𝐫𝐞𝐰𝐚𝐫𝐝 𝐟𝐨𝐫 𝐲𝐨𝐮𝐫 𝐰𝐫𝐢𝐭𝐢𝐧𝐠. fun fact. most people know @swyx as the famous podcaster, the AIE godfat

tutorialsswyx--x

26 Jun 2026

All entries

From Guessing to Placeholding: A Cost-Theoretic Framework for Uncertainty-Aware Code Completion

From Hallucination to Grounding: Diagnosing Visual Spatial Intelligence via CRISP

From Lexicon to AI: A Structured-Data Pipeline for Specialized Conversational Systems in Low-Resource Languages

From query to action: Introducing SQL alerting in Cloud Monitoring Observability Analytics

From Structure to Synergy: A Survey of Vision-Language Perception Paradigm Evolution in Multimodal Large Language Models

from tokenmaxxing to this in less than three months. anyone who doesn’t see this as affecting Anthropic and OpenAI is not paying attention.

From Vajrayana Tara to Bengali Baul: A Computational Study of Lexical Transmission Across Buddhist, Shakta, and Vaishnava Traditions in Bengal

From Weights to Features: SAE-Guided Activation Regularization for LLM Continual Learning

from Write & Learn (thx @MilksandMatcha @philipkiely @swyx), my most important takeaway, 𝐛𝐮𝐢𝐥𝐝 𝐚 𝐝𝐞𝐧𝐬𝐞 𝐫𝐞𝐰𝐚𝐫𝐝 𝐟𝐨𝐫 𝐲𝐨𝐮…

Frontier models are getting really expensive Model providers give us a bunch of features to reduce costs, but the support landscape is….inco…

fTNN: a tensor neural network for fractional PDEs

Full duration single-engine static fire test of Starship

Full spectrum Unlearnable Examples via Spectral Equalization

Fun fact: Dario once delayed the release of GPT-2 back at OpenAI, claiming it was too dangerous

“Gary Marcus … says that investments into scaling AI is the “greatest capital misallocation in history” which leaves everyone “on the hook”.…

@GaryMarcus, great interview and explanations of the challenges in AI. https://youtu.be/_pSivPlRx5o?is=dlFN6YBXsMA-tW4h

GAVEL: Grounded Caption Error Verification and Localization

GeMoE: Gating Entropy is All You Need for Uncertainty-aware Adaptive Routing in MoE-based Large Vision-Language Models

Generating a Paracosm for Training-Free Zero-Shot Composed Image Retrieval

Generating Special Triangulations with Transformers

Generative AI and Copyright Infringement: A Legal-Technical Analysis of AI Music Generation Systems Under 17 U.S.C. Title 17

Generative Models on Analog Hardware with Dynamics

Generative Retrieval via Diffusion Transformer with Metric-Ordered Sequence Training and Hybrid-Policy Preference Optimization

GenRecal: Generation after Recalibration from Large to Small Vision-Language Models

GEOALIGN: Geometric Rollout Curation for Robust LLM Reinforcement Learning

Geometric Fairness-Aware Routing for Federated Edge Networks

Geometric Gradient Rectification for Safe Open-Set Semi-Supervised Learning

Geometry-Aware MCTS for Extremal Problems in Combinatorial Geometry

Geometry-Aware Superpixel Graph Transformer with Metadata for Skin Lesion Classification

Goal! ⚽️🇯🇵🚀

gone camping @sequoia 🏕️ 🌙

Good new first: Sol is a smart, efficient, and a significant step forward. It is the same price as GPT-5.5. Also launching in the GPT-5.6 fa…

Good way to frame it. Also, if even one person had died, we would know who they were! Not one person has died from stopping waste & fraud at…

Governing Actions, Not Agents: Institutional Attestation as a Governance Model for Autonomous AI Systems

GPT-5.6 Sol is our most capable model yet for cybersecurity. It shifts the performance-efficiency frontier for long-horizon security tasks i…

GPT-5.6 Sol matches Mythos Preview on ExploitBench, adds Ultra mode with subagents for complex workflows, and max reasoning for deep problem-solving (OpenAI)

GPT‑5.6 Sol launches with our most robust safety stack yet. We strengthened real-time protections against high-risk cyber activity and repea…

Gradient Testing and Estimation by Comparisons

Graph Neural Networks Applications Across Domains: All Insights You Need

Graph Reinforcement Learning for Calibration-Aware Quantum Circuit Routing

Great experiment testing how good AIs are getting at very ambitious end-to-end coding tasks. Opus 4.7, in 14 hours, was able to build a soft…

Great to see the new GPT-5.6 models finally announced. Sad to see this new release strategy where only a select few get access initially. No…

Grok is balanced

Hallucination in World Models is Predictable and Preventable

Hardware Design for Table Tennis Robot Capable of Beating Professional Players

HarmVideoBench: Benchmarking Harmful Video Understanding in Large Multimodal Models

HauntAttack: When Attack Follows Reasoning as a Shadow

Have been taking different local open-weight LLMs for a test drive in different harnesses (Qwen-Code, Codex, Claude Code). 30B Mixture-of-Ex…