Notices by Srevin Saju (srevinsaju@mstdn.io), page 2

The deadline ended on Sunday, and my exams ended on Saturday. I really needed a break - let's try submitting something new!

Step 2: Keywords - Now that I have the skeleton concept ready, here are some keywords I gathered - #ai #puzzle #human #technology #collaboration and yes, #inkscape

Let's try to make AI and Humans work together on creating a new art style - maybe like, a missing piece of puzzle that an artist needs to put into their drawing perhaps.

Step 3: Mapping - Let's map the ideas out there on the sheet: I have this small idea where we have a robotic hand with a puzzle piece, trying to complete a picture, inkscape logo at the center, a human hand drawing with a paintbrush. The drawing divided into two halves, left side has a more technical art style, right side is more creative and natural.

I don't know how to draw a robotic hand holding something, and it has to be a puzzle piece. Let's try something on DALLE-2

Next, I had the idea where I would transition color swatches on the left side (technical) to a cityscape over different colored skies on the right (natural). Let's try this prompt: "Color swatches and transition to night sky digital art"

Although in both cases, I did not receive what I wanted as-is. But, it provided a good base case for me to work on, time to draw!

On to the blueprint!

#inkscape #splash #screen #about #art #mastoart #drawing #timelapse

As you can see, it looked very bad in the beginning, but I think it got better over time. This is a draft, and yes, I had to remove certain parts of this blue print too for the final picture.

@smoldesu @codewiz right! looks like the time taken grows exponentially with the size of the model.

@codewiz recently, I gave my shot with the Pygmalion AI 6B model, the results were not as bad as I thought it would be. I tried running it on Google Collab, and it worked pretty okay. Testing it locally is going to be a pain, since my laptop doesn’t have a GPU. The model was able to load and generate dialogues in 10 to 15 seconds, at a rate of 1 second per token, at 8 GB GPU RAM, 3 GB RAM and optimal CPU usage.

@codewiz I am not sure they understand the implications of the same, or how powerful ChatGPT3 is, but looks like the students have won a mini lottery.

@codewiz some students from the class tried using ChatGPT to solve the question, but apparently it didnt work out right. Here is a fun chat-

@codewiz Our CS professor apparently allowed us to use ChatGPT3 for our tests

@codewiz Ah names are really confusing, most of my professors and books refers to them as “procedures” :/

@codewiz I am using the nasm assembler, it was the one specified in our university study material.

I had been looking at what `MOV AH, 0x0E` means and I found this https://forum.osdev.org/viewtopic.php?f=13&t=31031, a 13 year old had been trying to make a bootloader. wow I feel so old.

Meanwhile [ORG 0x7c00] also seems to be born out of convention. The BIOS loads the program to the RAM at 0000:7C00 before your program gets a chance to be executed. We _could_ ignore the ORG (aka origin), but your program will assume that its loaded at 0x0000 instead. (But you could choose to add 0x7c00 on every command if you would like to do the hardway).

And I guess I am done, (at least with printing the character 'A').

#bootloader #boot #assembly

Looks like we have to choose "Floppy" storage type and choose the boot.img which we created earlier.

For our #assignment, we _have_ to use #VirtualBox to run them, and create screenshots of the output of the same.