Software

The conclusions of the post on Phi-4 left me stunned. How was it possible that a model like Phi-4 Reasoning Plus, which boasts an impressive 14.7 billion 4-bit parameters and was trained on scientific problems, particularly in mathematics, could have failed so badly?

When was the latest truly memorable WWDC? I’d say in 2020, a year that was already memorable in itself, when Apple unveiled the new Macs with Apple Silicon processors, capable of outperforming their equivalent Intel-based models.

After two long months, I was once again able to play again with LM Studio, and this post was supposed to provide a live description of the responses of some models I had just installed. However, things got out of hand when the first model I put under the magnifying glass, Microsoft’s 4-bit Phi-4, started behaving in strange ways that were worth describing in detail. From that moment on, the post you’re about to read practically wrote itself!

As I was writing about my transition from WordPress to Jekyll, I knew I had to prepare for another change. From a technical point of view, Jekyll is a fantastic platform: it is easy to program, has impeccable documentation, and works perfectly during the development phase, with a limited number of pages and test posts. But, as I experienced firsthand, when Jekyll is asked to handle a real site with hundreds of posts, performance drops dramatically and response times become unbearably slow (and quite embarrassing, too).

– Source: Markus Winkler on Unsplash. In the previous post I introduced the LM Studio interface, then tried the default suggested model (DeepSeek 7B) with one of the example prompts.

– Paul Allen sitting at the teletype connected to the school computer, with a very young Bill Gates standing next to him. Source: Celebrate 50 Years of Microsoft

Whether you Like it or not, Microsoft is the world’s largest software maker that, through methods both fair and questionable, has managed to get its products installed on the vast majority of computers on the planet.

As some of you may already know, I use LLMs (Large Language Models) for what they’re really good at, but I’m pretty skeptical about whether they’re truly intelligent or can solve any problem, as the folks at OpenAI, Microsoft, Google, and Meta keep telling us every day. They’ve invested a ton of money in LLMs, and they obviously have a big stake in getting everyone to use them all the time.

– Source: Scott Rodgerson on Unsplash. As promised (or threatened?) in the last post, this post is dedicated to exploring the available options for hosting a Jekyll-based website (or any other static site generator), whether for a personal blog like the one you’re reading, a professional studio, or a small business website.

As noted in the last post, Jekyll does not have a built-in commenting system, but its themes often allow comments to be managed by connecting to third-party services such as Disqus and similar platforms.