renzume.

Recent bookmarks

Y Combinator Supports AI Startup Dehumanizing Factory Workers

A Y Combinator-backed startup, Optifye.ai, has developed an AI surveillance system that monitors factory workers' movements and productivity through computer vision. The system, created by Duke University students from manufacturing families, allows supervisors to track worker efficiency in real-time and confront underperforming employees directly. The technology raises concerns about worker privacy and workplace conditions, similar to existing surveillance systems in remote work and Amazon warehouses.

EdgeDB is now Gel and Postgres is the Future | Gel Blog

PostgreSQL's unique position as a truly open-source, advanced relational database with extensive modularity is driving its accelerating adoption across the industry. EdgeDB announces its rebranding as Gel, introducing full SQL support and deeper ecosystem integration while maintaining its role as a PostgreSQL frontend layer enhancement. The platform aims to provide seamless database integration similar to how TypeScript enhances JavaScript.

Y Combinator deletes posts after a startup's demo goes viral | TechCrunch

An AI-powered worker monitoring system by Y Combinator startup Optifye.ai sparked controversy after demonstrating real-time factory worker surveillance capabilities. The demo video showing performance tracking of 'Number 17' led to social media backlash, resulting in YC removing the content from their platforms. The incident highlights growing concerns about workplace AI surveillance despite continued VC investment in similar technologies.

Gleam, coming from Erlang

A developer shares their experience building a feed aggregator using Gleam, a type-safe language running on the Erlang VM. The article explores Gleam's features, including its type system, error handling, and OTP integration, while highlighting both strengths and challenges in implementing a real-world application.

GitHub - deepseek-ai/DeepGEMM: DeepGEMM: clean and efficient FP8 GEMM kernels with fine-grained scaling

DeepGEMM is a CUDA library offering efficient FP8 matrix multiplications with fine-grained scaling, supporting both normal and Mix-of-Experts GEMMs. The lightweight library matches or exceeds performance of expert-tuned libraries, featuring runtime compilation and Hopper tensor core optimization, while maintaining a simple ~300-line core kernel.

New Maps of the Bizarre, Chaotic Space-Time Inside Black Holes | Quanta Magazine

Physicists are exploring the chaotic nature of space-time near black hole singularities using new mathematical tools and the AdS/CFT correspondence. Recent research revisits the Mixmaster universe theory from the 1960s, suggesting that space near singularities experiences chaotic stretching and compression in multiple directions. These studies aim to reconcile general relativity with quantum mechanics and potentially reveal fundamental truths about space-time.

Other kinds of talks

A critical analysis of common conference talk patterns suggests alternatives like distillation of complex knowledge, adversarial collaboration, replication studies, and failure analysis. The piece advocates for more substantive presentations that systematize knowledge, verify claims, and share valuable lessons from failures rather than superficial project updates or product pitches.

The XB-70

A personal account details the development of the XB-70 Valkyrie supersonic bomber through the lens of a father's career at General Electric during the Cold War. The ambitious aircraft project, despite its technological advances, was ultimately reduced from a strategic bomber program to a research vehicle due to the rise of ballistic missiles. The narrative weaves together family history, technological innovation, and Cold War politics while following the father's professional evolution from farm boy to engineer.

A Defense of Weird Research—Asterisk

Federal funding of scientific research, including seemingly obscure projects, consistently yields high returns with estimates suggesting $14-70 in social benefits per dollar spent. Government-funded basic research often leads to unexpected breakthroughs, like CRISPR and ARPANET, taking approximately 20 years from discovery to practical application. Private sector cannot effectively fund basic research due to long timelines and inability to capture full value, making government funding crucial for scientific advancement.

Hard problems that reduce to document ranking

Language models can effectively perform listwise document ranking, particularly useful in identifying N-day vulnerabilities through patch diffing. The technique transforms complex security problems into document ranking tasks, demonstrated successfully in locating vulnerable functions among patch diffs using GPT-4 mini with minimal cost and time.

Hello from Chicory | Chicory

Chicory is a JVM native WebAssembly runtime implemented entirely in pure Java, requiring no system resources. The runtime offers easy integration capabilities for plugin systems while maintaining security through sandboxed execution of WebAssembly modules.

Dogs may have domesticated themselves because they really liked snacks, model suggests

Scientists propose that wolves may have domesticated themselves into dogs by choosing to stay near humans for food and selecting mates with similar tameness preferences. A new statistical model demonstrates this self-domestication could have occurred over 15,000 years through natural selection, challenging previous assumptions about human-driven domestication.

Closing A Back Door In Illinois FOIA

A legal battle over database schema access through FOIA requests in Illinois reached the state Supreme Court, ultimately ruling that schemas could be withheld as 'file layouts.' New legislation (SB0226) aims to require public bodies to provide database structure descriptions, enabling citizens to request specific database queries.

'Hey Number 17!'

Optifye.ai, a Y Combinator-backed startup founded by Duke University students, introduces AI-powered surveillance system for monitoring factory workers' productivity through machine vision tracking. The system allows supervisors to monitor workers' hand movements and efficiency metrics in real-time, raising concerns about worker privacy and workplace conditions. Y Combinator has since removed promotional posts about the company's launch.

Introducing DeepSearcher: A Local Open Source Deep Research

DeepSearcher is an open-source research agent that builds upon previous work by adding features like conditional execution flow, query routing, and improved interfaces. The system leverages SambaNova's custom hardware for faster inference with the DeepSeek-R1 model, demonstrating advanced concepts in AI research automation through a four-step process of question definition, research, analysis, and synthesis.

More