Source: Hacker News
Source: NYT > Technology
The San Francisco start-up claims that its Chinese rival may have used data generated by OpenAI technologies to build new systems.
Source: Hacker News
Source: Ars Technica
On Monday morning, Nvidia stock dove 11 percent amid worries over the rise of Chinese AI company DeepSeek, whose R1 reasoning model stunned industry observers last week by challenging American AI supremacy with a low-cost, freely available AI model, and whose AI assistant app jumped to the top of the iPhone App Store's "Free Apps" category over the weekend, overtaking ChatGPT.
The drama started around January 20 when Chinese AI startup DeepSeek announced R1, a new simulated reasoning (SR) model that it claimed could match OpenAI's o1 in reasoning benchmarks. Like o1, R1 is trained to work through a simulated chain of thought process before providing an answer, which can potentially improve the accuracy or usefulness of the AI models' outputs for some types of questions posed by the user.
That first part wasn't too surprising since other AI companies like Google are hot on the heels of OpenAI with their own simulated reasoning models. In addition, OpenAI itself has announced an upcoming SR model (dubbed "o3") that can surpass o1 in performance.
Source: Hacker News
Source: Hacker News
Source: Hacker News
Source: Hacker News
Source: Ars Technica
Parents, students, teachers, and administrators throughout North America are smarting from what could be the biggest data breach of 2025: an intrusion into the network of a cloud-based service storing detailed data of millions of pupils and school personnel.
The hack, which came to light earlier this month, hit PowerSchool, a Folsom, California, firm that provides cloud-based software to some 16,000 K–12 schools worldwide. The schools serve 60 million students and employ an unknown number of teachers. Besides providing software for administration, grades, and other functions, PowerSchool stores personal data for students and teachers, with much of that data including Social Security numbers, medical information, and home addresses.
On January 7, PowerSchool revealed that it had experienced a network intrusion two weeks earlier that resulted in the “unauthorized exportation of personal information” customers stored in PowerSchool’s Student Information System (SIS) through PowerSource, a customer support portal. Information stolen included individuals’ names, contact information, dates of birth, medical alert information, Social Security Numbers, and unspecified “other related information.”
Source: Hack a Day
Back in November we first brought you word of a slicing technique by which the final strength of 3D printed parts could be considerably improved by adjusting the first layer height of each wall so that subsequent layers would interlock like bricks. It was relatively easy to implement, didn’t require anything special on the printer to accomplish, and testing showed it was effective enough to pursue further. Unfortunately, there was some patent concerns, and it seemed like nobody wanted to be the first to step up and actually implement the feature.
Well, as of today, [Roman Tenger] has decided to answer the call. As explained in the announcement video below, the company that currently holds the US patent for this tech hasn’t filed a European counterpart, so he feels he’s in a fairly safe spot compared to other creators in the community. We salute his bravery, and wish him nothing but the best of luck should any lawyer come knocking.
So how does it work? Right now the script supports PrusaSlicer and OrcaSlicer, and the installation is the same in both cases — just download the Python file, and go into your slicer’s settings under “Post-Processing Scripts” and enter in its path. As of right now you’ll have to provide the target layer height as an option to the script, but we’re willing to bet that’s going to be one of the first things that gets improved as the community starts sending in pull requests for the GPL v3 licensed script.
There was a lot of interest in this technique when we covered it last, and we’re very excited to see an open source implementation break cover. Now that it’s out in the wild, we’d love to hear about it in the comments if you try it out.
Thanks to [greg_bear] in the Hackaday Discord for the tip.
