Just make Sidebery one of those 3 or 4 recommended extensions when you first install Firefox. That extension is fantastic. Typically I'll disable the internal tabbing system entirely and only use Sidebery with auto-hiding.
So, can a distilled 8B model (say, the Deepseek-R1-Distil-Llama-8B or whatever) be "trained up" to a higher parameter 16B Parameter model after distillation from a superior model, or is it forever stuck at the 8B parameters that can just be fine tuned?
Can't wait til Jan 2025. I'll be taking full advantage of the home energy rebates to rewire the electric, install modern insulation, and install energy efficient heating/cooling in my 1930's Craftsman. The rebates should cover up-to $14,000 of that cost which will be a great help.
Interesting, I didn't read that as a DEI hire at all. I interpreted it as though he was just hinting at or throwing out a clue as to who he had already selected.
He took 2012-2013, and 2016-2021 off work to travel? That large of a gap doesn't look great on a resume. He basically worked half or less of the 2010s? Ridiculous.
The more things like this happen, the more the function of government (and business, and relationship between labor and business) will return to the way they were operated in the US between 1880 and 1920.
Are you a citizen of China, or move there for work/education/research?
Anyway, this is very unrelated, but I'm in the USA and have been trying to sign up for the official learning center for CAXA 3D Solid Modeling (I believe it's the same program as IronCAD, but CAXA 3D in China seems to have 1000x more educational videos and Training on the software) and I can't for the life of me figure out how to get the WeChat/SMS login system they use to work to be able to access the training videos. Is it just impossible for a USA phone number to receive direct SMS website messages from a mainland China website to establish accounts? Seems like every website uses SMS message verification instead of letting me sign up with an email.
Wow, I figured this would have been the most basic level of common knowledge for LLM end-users by this point. I guess there is still a surprising amount of people out there who haven't jumped on the bandwagon.
I'm honestly at the point where I wish there were a built-in one-click filter for Hacker News to prevent anything related to Machine Learning, AI, Chatbots, etc, from showing up in the feeds here when I'm signed in.
Even better if it were as easy and intuitive to use as IronCAD. It's by-far the best UI/UX CAD software and I'll never understand why it doesn't get more recognition.
They must work for one of these companies. The last 6 months of their post history is a majority of just showing up when an article like this appears and defending PFAS and other types of chemicals or poisons.
That's what I immediately understood the title to mean... Is this not what the title is saying? I'm confused as to why people are having trouble understanding it.
Yes, all the possible length extending custom instructions you can think of. I can get some reasonable length responses out of it, but I've never seen them go over 1 page worth, and multi-shot example prompts using multiple USER and GPT exchanges to define the format. Seems like GPT4 has a hard limit as to how much it will output when you click "continue", and Claude Opus never goes over a page either. Another user pointed out using the API, which I have done in the past, but it's been a long while, and I can't really justify the cost of using the advanced models via API for my general use.
Long Context is great and all, but it sucks that all of these LLM's have really poor output length. If I feed something an entire book and ask for a comprehensive summary then I'm expecting at least a full 3-page summary. I get that they try to force these things to be "concise" to save on compute, but good lord it's so annoying.