Daring Bits
Americans Out, H-1B In - Major tech firms are under fire for hiring foreign workers via the H-1B visa program as they lay off hundreds of thousands of American employees. Republican-backed reforms would set wage floors to mitigate the situation, though when combined with the recent tax changes this combination could be a fatal blow to the US software industry.
Open Minds, Rotten Flesh - openSUSE is “Linux for open minds,” so long as those open minds aren’t conservative. Otherwise, you’re “rotten flesh,” according to the unrepentant openSUSE and SUSE leadership.
Open Source Laptop - Lukas Henkel’s is on a mission to decode and democratize the realm of high-tech consumer electronics. Promising a goldmine of educational content, the project aims to dismantle the ‘black box’ mystery of device construction, inspire sustainable and repairable design choices, and overcome complex multidisciplinary challenges in an effort to bring everyone into the future of personalized, open-source laptops.
Build a Modern Computer from First Principles: From Nand to Tetris - This highly rated Coursera course elucidates the fundamentals of how computers work. Also, Tetris.
KryptEY - Here’s an Android keyboard for secure E2EE communication through the signal protocol in any messenger. Communicate securely and independent, regardless of the legal situation or whether messengers use E2EE
Robot (Government) Overlords
Employable Humanity - Frederick M. Hess examines which career clusters are most vulnerable to AI-automation, and proposes that a blend of technical and transferable skills can potentially create a workforce that is resilient to AI threats, ensuring sustainable career paths in the face of emerging technologies.
Why can ChatGPT talk? - Stephen Wolfram discusses at length the inner workings of ChatGPT and where it gets its powers, concluding that it has implicitly discovered regularities in human language. If these could be defined explicitly, it may open the door to doing the ChatGPT magic in vastly more direct, efficient—and transparent—ways.
The Hard Things About LLMs - Phillip Carter at honeycomb.io says Large Language Models (LLM) aren’t magical and can’t solve every problem for you. In this article, he walks through a number of the inherent challenges involved in building products with LLMs.
Bogus Briefs - Lawyer Steven Schwartz faces potential sanctions after confessing to the use of ChatGPT to compose court documents, which included references to six non-existent cases fabricated by the chatbot. Who didn’t see this coming?
Humans in 4D
Mult-Agent Debate
What indeed can GPT models do in chemistry?
Bertopic
Large Disagreement Modeling
