I continue working on my side project Babel Nexus. I had some issues with billing so I moved it to my home computer and am rebuilding it using powershell and VS Code. I am trying to figure which small AI system I should put on my computer like Cline, Zoo Code, or Qwen. I can do this because I have an NVIDIA graphics card. I have also looked at Ollama. I am steadily rebuilding my app. It costs very little to build it on my home computer. I have redone a lot of little things and added a whole bunch of content. I plan on using Railway to host it because for the long term it is less expensive. Right now, my existing app is placeholder. https://babelnexus.com I added several new free ebook applications recently to the Beyond the Public Domain section.
There are other things which might make more sense without strings attached in the same way. If you can come up with a research idea, there is the SBIR Small Business Investment Research grant which can be combined with Innovation Grants. In NY, this is the NY Star Innovation Grant. My understanding is that about 20% who apply get them. In a way this is as good or better odds of getting funding than VC money if you are high tech company with a patentable idea and want to start. Plus there are other grants and funding which are more local.
I read everything I can get my hands on, sometimes even things I don't understand at all. Occasionally you have to stretch yourself and look at things that don't make sense. Part of working with LLMs is understanding how the programs works, this is different than writing the code. The explanations coming from the LLM have to make sense to you, or you will get garbage or not understand what your applications is actually doing. You have to be able to look at the code and say I understand what this is doing even if you can't write it. A good example of this is if there is a product manager running a team. They have to understand what the code does even if they don't write it.
Wikipedia is going to change with this. I expect more apps and video instruction to happen. There is no reason Wikipedia cannot expand Wikimedia and make it more effective. They have a huge amount of creative commons content already on Wikimedia. Wikimedia is already better than Wikipedia in many ways.
What is going on is you get a lot of people who pretend or actually have domain expertise who want to make something. The problem is that there is no real guarantee of actually having the expertise or taking the time to make something that is decent. Thus you get someone who either knows a lot about wine for example or has a hobby about wine who wants to make a wine site. They then try to make a wine site and depending on their motivation make something terrible or if they are obsessed something which is somewhat decent after a lot of effort. The other problem is that AI is very hard to make anything above B grade material. You need a technical person to help to make a commercial grade site. Thus you get a lot of mediocre sites, some truly awful sites, and a few slightly above average sites. The other thing that goes with this is that there is no control over the veracity, accuracy, normalcy, or radicalness of the sites being created. It is turly anything goes. People want to show what they are interested in or their expertise...
Anyone with an IQ over 110 can make a commercial grade grade site these days if they are generally curious and can ask good questions. Opus 5.5 + Cloudflare + Supabase/Neon/etc. will get you to thousands of users no problem. By the time enterprise scale becomes a problem, Opus 6.5 will solve it.
I'm not fully confident that they truly want to show what they are interested in, even if they don't have deep expertise in it. My feeling is that there is a notable portion of projects being the result of prompting a chat-bot for ideas to develop. ChatGPT, what I can do? Make a website about wine. :)
It might be useful on Mars where there is lower gravity, especially with thrusters. You might be able to go much farther. It is after all where Musk might go.
Pretty much all of the coding LLMs, Claude, ChatGPT, etc. can generate a product demo. I had Replit make my product demo. For a personal presentation, Loom allows you to edit your presentations wordings directly for that two minute video of yourself which you make from your computer monitor.
Interesting!
Was Replit after the coding LLMs? Any advantage or Replit over them?
Anything you'd like to have but not yet delivered by the current solutions?
reply