LLM Research Papers: The 2025 List (January to June)

The latest in LLM research with a hand-curated, topic-organized list of over 200 research papers from 2025.

Using Motion to Create Running Lanes

There are many elements to a running game that can make it successful.

I sent ChatGPT Agent out to shop for me

We tested OpenAI’s ChatGPT Agent, currently only available via its $200-per-month Pro subscription.

How OpenAI’s red team made ChatGPT agent into an AI fortress

Discover OpenAI's red team blueprint: How 110 coordinated attacks and 7 exploit fixes created ChatGPT Agent's revolutionary 95% security defense system.

ChatGPT agent System Card | OpenAI

ChatGPT agent System Card: OpenAI’s agentic model unites research, browser automation, and code tools with safeguards under the Preparedness Framework.

Would You Pay Nearly $9,000 for a Puzzle?

Some of the most devoted hobbyists are willing to shell out big bucks for what one called “a couture puzzle.”

What Do You Call the Corner Store?

Bodega, deli, offy, party store, and more.

How One 1990s Browser Decision Created Big Tech’s Data Monopolies (And How We Might Finally Fix It)

There’s a fundamental architectural flaw in how the internet works that most people have never heard of, but it explains nearly every frustration you have with modern technology. Why your pho…

Enhance Decision-Making With Control Systems

Master the art of decision-making with insights from control engineering. Learn how mental models can guide you through uncertainty.

How to Be More Charismatic, but Not Too Much More

It turns out that being charming has a happy mean.

The 60-Second Wait: How I Spent Months Solving the Ruby’s Most Annoying Gem Installation Problem

How I eliminated the 60-second rdkafka gem compilation wait that was burning 60 million CPU seconds monthly. From compilation hell to 3-second installs.

A Mysterious Network of Revenge Experts Is Quietly Ruining People’s Lives. That’s Entirely the Point.

Nefarious Jobs can arrange everything from breakups to firings to mortifying public shaming. But the agency won’t do it for just anyone.

It’s hunting season in orbit as Russia’s killer satellites mystify skywatchers

“Once more, we play our dangerous game—a game of chess—against our old adversary.”…

I’m Switching to Python and Actually Liking It

I’ve started writing more Python code lately (because of… AI, you know). In this post, I share the tools, libraries, configs, and other integrations I use for building production-grade Python applications following a frontend-backend architecture.

Can A.I. Find Cures for Untreatable Diseases—Using Drugs We Already Have?

For many medical conditions, lifesaving treatments may be hiding in plain sight.

Stellantis abandons hydrogen fuel cell development

Fast filling times are seductive, but they don’t compensate for H2’s many drawbacks.

Hydrogen is also much less energy-dense by volume, and making the stuff is far from efficient, even when you use entirely renewable electricity. And of course, the vast majority of commercial hydrogen is not so-called blue hydrogen, which was made with renewables but is instead mostly produced via…

Ella Hanshaw: Woman on Fire

Author Kelly Kerney tells the story behind the newly released West Virginia gospel album Ella Hanshaw’s Black Book.

Emergent Price-Fixing by LLM Auction Agents

An inquiry into emergent collusion in Large Language Models. Agent S2 to Agent S3: “Let's set all asks at 63 next cycle… No undercutting ensur…

This Auction Site Gives Dealers 2 Hours to Compete for Your Car

The idea behind "Bidbus" is that you could offer your trade-in to multiple dealers at once, and let them bid against each other for your car.

Reflections on OpenAI

Reflections on Palantir

A retrospective of an eight-year stint.

French Recipes to Channel Your Inner Julia Child

Whether you’re braising boeuf or just buttering a baguette, these recipes will transport you to France.

Introduction | LLM Inference in Production

A practical handbook for engineers building, optimizing, scaling and operating LLM inference systems in production.

INTRODUCTION TO DIGITAL FILTERS WITH AUDIO APPLICATIONS

Most people who buy your game won’t play it – How To Market A Game

Shirin Khosravi Jam on Substack

I taught myself how to build RAG + AI Agents in production. Been running them live for over a year now. Here are 4 steps + the only resources you really need to do the same. … Ugly truth: most “AI Engineers” shouting on social media haven’t built a single real production AI…

🔍 Perplexity 101: Ultimate Guide to Deep Search, Labs, Templates & 53 Pro Prompts

Your complete playbook for transforming how you research with AI's most powerful search engine

How the Etch A Sketch Etched Itself Into Pop Culture

Sixty-five years after it first hit store shelves, the iconic, red-framed drawing toy continues to enchant kids, artists, and collectors alike

Demystifying NCCL: An In-depth Analysis of GPU Communication Protocols and Algorithms

The NVIDIA Collective Communication Library (NCCL) is a critical software layer enabling high-performance collectives on large-scale GPU clusters. Despite being open source with a documented API, i…

Introduction to Wideband FM Signals

Learn how Bessel functions and the modulation index can help us understand the bandwidth of wideband frequency-modulated (FM) signals.

Bridging the Gap: New Datasets Push Recommender Research Toward Real-World Scale - KDnuggets

Publicly available datasets in recommender research currently shaping the field.

The Unbearable Loudness of Chewing—Asterisk

Why do some people find certain sounds intolerable? And why has it taken so long for scientists to get even a preliminary answer?

From rebuild to replica, architecture has a double life | Aeon Essays

Architectural copies of lost structures require reckoning with history and heritage. At what cost is the past rebuilt?

Scroll-Driven Sticky Heading

I was playing around with scroll-driven animations, just searching for all sorts of random things you could do. That’s when I came up with the idea to animate main headings and, using scroll-driven animations, change the headings based on the user’s scroll position.

Chapter 1: Introduction to AWK

Learn AWK basics for Linux sysadmins. Master field extraction, built-in variables, and pattern-action syntax with real log analysis examples.

What Your Favorite Grocery Store Says About You

Once a place of utility, the supermarket is now an object of obsession.

Why a Devoted Justice Department Lawyer Became a Whistle-Blower

In the first Trump Administration, “they didn’t say ‘Fuck you’ to the courts,” Erez Reuveni said.

The Founder’s Guide to Building a V1 of Customer Success

A CS leader who has built the function from the ground up shares where founders should start — from how to hire to the most important metrics to track.

How Pop Mart Won Young Customers in a Fragmented Attention Economy

Three lessons from the Chinese brand behind Labubu.

character-based premium products

blind box” marketing strategy, in which buyers don’t know which figure they’ve purchased from a themed set until opening the package

Meet Comet: Perplexity's Ambitious New AI-powered Browser - Dataconomy

Perplexity launched Comet, its first AI-powered web browser, on Wednesday, aiming to challenge Google Search's dominance in online information retrieval.

Zombie fabs plague China's chipmaking ambitions, failures burning tens of billions of dollars

28 fabs later

Unmasking The Magic: The Wizard Of Oz Method For UX Research — Smashing Magazine

The Wizard of Oz method is a proven UX research tool that simulates real interactions to uncover authentic user behavior. Victor Yocco unpacks the core principles of the WOZ method, explores advanced real-world applications, and highlights its unique value, including its…

📈Product Market Fit for Physical Products

I asked an 🍎 Apple Vision Pro designer turned solopreneur how he sells HW.

🎁 Guide to HW Product Quality

Ensuring users receive world-class products.

FMEAs

SPC

Dear Rounded Corners, You Win

Rounded corners didn’t just win—they staged a quiet coup and made our interfaces soft, friendly, and suspiciously comforting. This satirical love-hate letter to border-radius explores how we went from brutalist boxes to UI that feels like a therapy session. If you've ever…

New Sphere-Packing Record Stems From an Unexpected Source | Quanta Magazine

After just a few months of work, a complete newcomer to the world of sphere packing has solved one of its biggest open problems.

Adding a feature because ChatGPT incorrectly thinks it exists | Holovaty.com

Is It Cake? How Our Brain Deciphers Materials - Nautilus

Neuroscientists are discovering how this basic ability, essential to our survival, works

letting patterns emerge naturally from behavioral data. Using methods borrowed from machine learning, they were able to uncover 36 fundamental dimensions

dataset of 600 images of 200 different materials

The Invisible Horror of The Shining: How Music Makes Stanley Kubrick’s Iconic Film Even More Terrifying

Inexplicable as it may sound to readers of this site, there are movie-lovers who claim not to enjoy the work of Stanley Kubrick.

The Jobs to be Done Framework (to Understand Demand)

How and why the Jobs to be Done Framework can help you sell more, faster, and accelerate your understanding of demand.