1141
MCP: An (Accidentally) Universal Plugin System
Or: The Day My Toaster Started Taking Phone Calls
New
The latest blog posts to take off on Hacker News. 3,091 posts.
1142
AI Is Dehumanization Technology
A call to reject the deployment and use of AI systems in Canada's public sector.
1143
1144
OpenAI charges by the minute, so speed up your audio
I discovered a fun and strangely obvious trick for summarizing videos faster and reducing costs: just speed them up. Cheaper, faster OpenAI transcriptions with a little ffmpeg trick.
1145
Reinforcement learning, explained with a minimum of math and jargon
To create reliable agents, AI companies had to go beyond predicting the next token.
1146
The bitter lesson is coming for tokenization
Highlights the desire to replace tokenization with a general method that better leverages compute and data. We'll see tokenization's fragility and review the Byte Latent Transformer arch.
1147
TPU Deep Dive
I've been working with TPUs a lot recently and it's fun to see how they had such different design philosophies compared to GPUs.
1148
AI is going to hack Jira
Hide your kids, hide your wife: how the Agile Industrial Complex has set us all up to be duped by AI codegen
1149
Compiling LLMs into a MegaKernel: A path to low-latency inference
TL;DR: We developed a compiler that automatically transforms LLM inference into a single megakernel — a fused GPU kernel that performs…
1150
How OpenElections uses LLMs
In the 12-plus years that we’ve been turning official precinct election results into data at OpenElections , the single biggest problem has been converting pictures of results into CSV files. Many of the precinct results files we get are image PDFs, and for those there are essentially two options: data entry or…
1151
Andrej Karpathy's talk on the future of the industry
Transcript of Andrej Karpathy's YC AI SUS talk at Y Combinator on June 17th, 2024.
1152
Revisiting Minsky's Society of Mind in 2025
AI Agents, Modularity, and Alignment
1153
Building agents using streaming SQL queries
Table of Contents Agents Need to Interact With LLMs Agents Should Be Event-Driven Agents Need Context Agents Require Memory When SQL Is Not Enough Parting Thoughts AI Agents have improved in leaps …
1154
Is there a half-life for the success rates of AI agents?
Building on the recent empirical work of Kwa et al. (2025), I show that within their suite of research-engineering tasks the performance of AI agents on longer-duration tasks can be explained by an extremely simple mathematical model — a constant rate of failing during each minute a human would take
1155
I counted all of the yurts in Mongolia using machine learning
The Fall of Civilizations podcast put out a 6¾-hour episode on the history of the Mongol Empire, which I eagerly listened to. After finishing the episode I wondered about contemporary Mongolian society, I wanted to learn what the lands that the Mongol Empire exploded from are like in our current day. There are many…
1156
What Google Translate can tell us about vibecoding
There has been rather a lot of doomsaying (and perhaps astroturfing) lately about LLMs as the end of computer programming. Much of the discussion has been lacking nuance, so I’d like to add mine. I see claims from one side that “I used $LLM_SERVICE_PROVIDER to make a small throwaway tool, so all programmers will be…
1157
LLMs pose an interesting problem for DSL designers
The most exciting part of Programming Languages (PL) research for me has always been in Programming Language Design.
1158
O3 Turns Pro
You can now have o3 throw vastly more compute at a given problem.
1159
Claude Code feels like magic because it is iterative
Personal blog of Omar Abid
1160
Generative AI coding tools and agents do not work for me
People keep asking me If I use Generative AI tools for coding and what I think of them, so this is my effort to put my thoughts in writing, so that I can send people here instead of having to repeat…
1161
The drawbridges come up: the dream of a interconnected context ecosystem is over
The AI era is speedrunning the Web 2.0 story. Open and accessible MCPs are not our future. Integrations will be tightly governed.
1162
Snorting the AGI with Claude Code
blog - kade killary
1163
Meta's Llama 3.1 can recall 42 percent of the first Harry Potter book
New research could have big implications for copyright lawsuits against generative AI.
1164
Q-learning is not yet scalable
Over the past few years, we've seen that next-token prediction scales, denoising diffusion scales, contrastive learning scales, and so on, all the way to the point where we can train models with billions of parameters with a scalable objective that can eat up as much data as we can throw at it. Then, what about…
1165
Seven replies to the viral Apple reasoning paper and why they fall short
Also: another paper that seals the deal
1166
Design Patterns for Securing LLM Agents Against Prompt Injections
This new paper by 11 authors from organizations including IBM, Invariant Labs, ETH Zurich, Google and Microsoft is an excellent addition to the literature on prompt injection and LLM security. …
1167
Agentic Coding Recommendations
Current recommendations of agentic coding.
1168
The first big AI disaster is yet to happen
The first public passenger locomotive, Locomotion No. 1, began service in September 1825. The first mass-casualty railway disaster happened seventeen years…
1169
AI at Amazon: A case study of brittleness
A year ago, Mihail Eric wrote a blog post detailing his experiences working on AI inside Amazon: How Alexa Dropped the Ball on Being the Top Conversational System on the Planet. It’s a great …
1170
Fine-tuning LLMs is a waste of time
People think they can use Fine-Tune for Knowledge Injection. People are Wrong