🌸 Thu, Sep 5, 2024 🌸
Do you like to write Git commit messages? We should make an AI to do it. I am going to write that, and it should not take too long, maybe an hour.
🌸 Wed, Sep 4, 2024 🌸
What am I doing? Laying down in the bed and reading C++ code to entertain myself.
🌸 Wed, Sep 4, 2024 🌸
We should use domestic heat to measure temperature instead of foreign heat.
🌸 Wed, Sep 4, 2024 🌸
Taking a sick day today; long COVID sucks. I am so angry.
🌸 Tue, Sep 3, 2024 🌸
I need to figure out how to make viral videos; I see this as the only way to promote my games. I don’t have a budget for traditional marketing.
🌸 Tue, Sep 3, 2024 🌸
Excavators did replace shovels, but look at the construction site and you still see people with shovels. AI will replace a lot of human activities: writing, drawing, music, and you will still see people in the future with “shovels” manually adjusting the AI output.
🌸 Tue, Sep 3, 2024 🌸
Yesterday I talked to my Gen Z son and learned that there are no old or new memes; in general, memes are plain old. Nowadays, it’s all just meaningless alpha humor, something completely random, without any story or meaning, it’s just random.
🌸 Sun, Sep 1, 2024 🌸
Let me know if you need an invite to Deadlock.
🌸 Sun, Sep 1, 2024 🌸
Oracle, why did you break my deploy script?
Do you accept the Oracle No-Fee Terms and Conditions (NFTC) for Oracle Java SE terms? [yes/no]
I just want to play Minecraft.
🌸 Sun, Sep 1, 2024 🌸
What if we trained LLMs on Abstract Syntax Trees instead of word tokens, allowing them to learn the underlying structure of programming languages and generate more correct, efficient, and idiomatic code?
🌸 Sun, Sep 1, 2024 🌸
Let’s revisit the concept of Large Language Models (LLMs) and backtracking. I’d like to explore how we can leverage LLMs to generate code. By restricting the types of tokens the LLM can generate, we can force it to produce valid queries.
But why stop there? On top of verifying the correctness of the code, we can take it a step further by attempting to compile and run tests on the generated code. If the code fails, we can backtrack, revert the LLM’s state to a previous step, and try again. If all possibilities are exhausted or have a low score, we can backtrack further.
This process mimics real-world programming: try, test, refine, and repeat. By combining LLMs with backtracking, we can create a more robust code generation system.
Backtracking is not a new concept in computer science. It’s a technique used in various algorithms, such as constraint satisfaction problems, puzzle solvers, and optimization problems. However, one of the challenges with backtracking is finding effective heuristics to navigate the solution space.
This is where LLMs can shine. By incorporating the LLM’s sorted list of tokens into the existing heuristic, we can guide the backtracking process to explore more promising areas of the solution space first. This could lead to faster convergence to a solution or even discovering better solutions that would be difficult to find through traditional backtracking methods.
🌸 Sun, Sep 1, 2024 🌸
Just keep swimming.
🌸 Sun, Sep 1, 2024 🌸
llm with backtracking
Probably I need to resubscribe to OpenAI, as I found that Meta.AI is still less smart than ChatGPT and also has some missing features, such as modifying the prompt or uploading documents.
Here’s the idea: combine a relational database (I’m currently considering SQLite) with a Large Language Model (LLM), specifically Llama 3 with 8B parameters. The goal is to enable the LLM to:
- Create tables and populate them based on user input
- Perform queries on the database
- Interpret the database output and provide the results back to the user in natural language
Additionally, I’d like to explore the possibility of providing the LLM with an existing database schema, allowing it to convert human queries in English into SQL. This would enable users to interact with the database using natural language, making it more accessible and user-friendly. By having the LLM interpret the database output, it can provide the user with a clear and concise answer in natural language, rather than requiring the user to understand SQL or database terminology. This would make the system even more user-friendly and accessible to a wider range of people.
Some progress, today I was not struggling to remember the password. Hopefully, my declined cognitive abilities are not permanent. #longCOVID
I will rename Neat Corp to NeatAid Corp. A quick Google search did not show anything, hopefully it will stay like this.
If you’re low on neat power, have a NeatAid.
This is for my game in development, Ordy: The Rise of Clara, where an evil corporation harvests neat power from powered people and they become ordy (short for ordinary). Initially, I called that company Neat Corp, and yesterday I learned that company exists. So, I need to “rebrand” the company to avoid clashing with the existing one.
Music teachers are angels; you cringe if someone is off by a couple of cents in the pitch. Imagine it’s 99.9999% of the time, all the time, all day.
I just learned that Neat Corp is a real company, in fact, it’s a game studio that made Budget Cuts. I should have Googled it first before integrating it deeply into my game. Grrr…
Have you found tooltips in Unreal Engine annoying? You can disable them. Press ` and type Slate.EnableTooltips 0
Can we make glasses that automatically remove dust, cover themselves automatically if I try to touch them, and also protect themselves from the rain and fog?
Hey! I can smell coffee again, thank goodness it wasn’t permanent loss.
Just keep swimming.
Don’t give up on your dreams. Go back to sleep. 😂
My sense of smell has become selective. I cannot smell any coffee. I had a bitter, warm, dark drink instead of coffee, with zero smell. I put my nose directly in the cup, nothing.
Previous - 1 - 2 - 3 - 4 - 5 - 6 - 7 - 8 - 9 - 10 - 11 - 12 - 13 - 14 - 15 - 16 - 17 - 18 - 19 - 20 - 21 - 22 - 23 - 24 - 25 - 26 - 27 - 28 - 29 - 30 - 31 - 32 - 33 - 34 - 35 - 36 - 37 - 38 - 39 - 40 - 41 - 42 - 43 - 44 - 45 - 46 - 47 - 48 - 49 - 50 - 51 - 52 - Next