Current LLMs are manifestly different from Cortana (🤢) because they are actually somewhat intelligent. Microsoft’s copilot can do web search and perform basic tasks on the computer, and because of their exclusive contract with OpenAI they’re gonna have access to more advanced versions of GPT which will be able to do more high level control and automation on the desktop. It will 100% be useful for users to have this available, and I expect even Linux desktops will eventually add local LLM support (once consumer compute and the tech matures). It is not just glorified auto complete, it is actually fairly correlated with outputs of real human language cognition.

The main issue for me is that they get all the data you input and mine it for better models without your explicit consent. This isn’t an area where open source can catch up without significant capital in favor of it, so we have to hope Meta, Mistral and government funded projects give us what we need to have a competitor.

deleted by creator

This is another reminder that the anomalous magnetic moment of the muon was recalculated by two different groups using higher precision lattice QCD techniques and wasn’t found to be significantly different from the Brookhaven/Fermilab “discrepancy”. More work needs to be done to check for errors in the original and newer calculations, but it seems quite likely to me that this will ultimately confirm the standard model exactly as we know it and not provide any new insight or the existence of another force particle.

My hunch is that unknown particles like dark matter rely on a relatively simple extension of the standard model (e.g. supersymmetry, axioms, etc.) and the new physics out there that combines gravity and QM is something completely different from what we are currently working on and can’t be observed with current colliders or any other experiments on Earth.

So probably we will continue finding nothing interesting for quite some time until we can get a large ML model crunching every single possible model to check for fit on the data, and hopefully derive some better insight from there.

Though I’m not an expert and I’m talking out of my ass so take this all with a grain of salt.

Cool meme but Reuters doesn’t own AP and Rothschild doesn’t own Reuters. It is quite ironic to be pushing against the very real problem of media disinfo/government propaganda trickled down through AP/Reuters, while at the same time spreading misinformation.

This isn’t an actual problem. Can you train on post-ChatGPT internet text? No, but you can train on the pre-ChatGPT common crawls, the millions of conversations people have with the models and on audio, video and images. As we improve training techniques and model architectures, we will need even less of this data to train even more performant models.