Microsoft’s recent move to challenge Google with their new ChatGPT AI has been met with much interest and speculation. Could this technology be the tech coup of the century that Microsoft is hoping for? It certainly seems possible, especially given Google’s lackluster response so far. But while it is tempting to get caught up in the hype of a potential Microsoft victory, we must also consider how this challenge could affect our everyday lives.

The recent conversations about large language models, such as ChatGPT, and the potential of them being “Google killers” have brought to light some interesting questions. How will Google pivot its ad engine to work with these AI-driven chat systems? What does this mean for content creators? What are the economics of running these models? Will this disrupt Google or will they simply evolve and make more money? These questions have no easy answers. The training costs for large language models like GPT are enormous, and the inference costs are substantial too. Right now things like ChatGPT are very cool parlor tricks, but there’s absolutely no way to justify them in terms of the economics of running the service today. In addition, there is a lack of incentive for content creators - what value does ChatGPT give back to primary sources? Without an answer to this problem, it is difficult to see how ChatGPT can be economically viable in the long term.

Mick West’s breakdowns of the Pentagon UAP videos have been an eye-opening experience for many. Initially, these videos were thought to be solid proof of alien activity, leading to a groundswell of support from those who believed in UFOs and even U.S. Senators pushing for their declassification. However, when they were released to the public, Mick West was quick to point out that what was seen on the videos could easily be explained as sensor glare, bokeh or mis-understood perspective. Additionally, West highlighted how easy it is for complex sensors like FLIR (Forward Looking Infrared) and radar systems to misinterpret data – even if multiple pilots witnessed them with their eyes or multiple other sensors picked them up!

The memory footprint of modern software has grown exponentially in the last 20 years. With features such as HDR, fluid animations, monstrous resolutions, 3D elements, accessibility options and fancy APIs for easier development allowing for more features than ever before, it’s understandable why this is the case. However, while it’s necessary to have access to a large amount of RAM to run these programs properly, there are also some downsides. Developers are often too busy working on new features to have time allocated towards refactoring their code in order to make the software more efficient and use less RAM.

Recent news of Samsung’s 870 EVO, 980 Pro and 970 Evo Plus series SSDs failing within 6 months has put many users on edge. Not only have these devices failed at an unprecedented rate, they have been reported to cause uncorrectable read errors that can lead to data loss and system downtime. To make matters worse, Samsung has been slow in providing clear information or acknowledging the issue publicly.