kanemcgrath

I was able to run it in Skyrim on my 5060ti at about 60fps. First game that almost used all 16gb of vram

Or Eldin ring naming your horse "Torrent" so that searches for "Eldin Ring Torrent" would bring up your horse instead of pirate downloads

I find it unfortunate that tele-operated robotics could be a legitimate and successful strategy. But everyone is so desperate for AI hype money that I don’t think we will see it for a while.

I really think dwarf planet is a fine classification. Pluto is a planet, just a really small one. Same with Eris and the other planets in the Kuiper Belt. I don’t know why the planet count of our solar system is seemingly such a Religious number. It seems as if it is the focus of why we should define what a planet is.

When I think of what a planet is, I think of a destination. Somewhere that maybe in the far future we could visit one day. I think the definition for a planet should be a body with sufficient gravity that we could reasonably orbit and land on. And we can have as many sub-classifications as we want from nano-planet to Mega-planet.

Googles local gemma models which target roughly the same parameter count range, are known for being a lot better at vision tasks than qwen, no idea if 3.8 has changed that though

whats the tok/s you get on that. I have heard a few claims of around 30-50 with mtp, but for how cheap the setup is I am surprised I don't hear more about 3060 stacks so I assume there has to be some catch.

I think I am going to buy a second rtx 3060, as 27B has been just outside of my range for to long, and this looks like the parameter count tipping point

Because of all the svg rendering stuff, I added a draw_svg tool to my harness and it has been really nice to get a quick mock-up of ui changes. And conveniently, the new DeepSeek models are really good at knowing when to use it. So I do look at pelican rendering as a small metric of useful capability.

I tried it for a bit, and It was not really worth its size. It got swept up in all the other AI news recently, but laguna s 2.1 I think is the best ~100B moe model right now

My favorite one was the blender addon https://github.com/libsm64/libsm64-blender

Also the steam DRM on a lot of single player games can be removed very easily. https://github.com/atom0s/Steamless

and is a recommended step for engine modding games like skyrim.

One of my favorite things about my custom mechanical keyboard, is being able to remap the entire key set in the firmware with VIA. I have fn+arrow keys for media, fn+space for play/pause fn+end for calculator, and a bunch of random others. It is so useful I could never get another keyboard that doesn’t have a similar functionality.

Something about qwen models hold up really well even at low quants. for most other models anything under q5 is cooked, but on 35B-A3B I can get a lot of things done even at q3_xl. It is definitely better than full precision 9B

I am always worried that popular software like this is getting compromised. But what about all the smaller stuff? I download and run a bunch of small programs, especially for game modding. Do the small tools not have the same risk because of their smaller scale invites less dedicated attacks? or do they get compromised and fly completely under the radar?

Mechjeb does a lot of automation for complex maneuvers like planetary transfers, ascent autopilot, target interception, etc.. You can even do a porkchop plot for interplanetary transfer. but lagrange points are not possible in KSP because of the Sphere of Influence Model. and orbital slingshots are not included, I assume because they require an intense bruteforce calculation.

I relied on Mechjeb too much in my game, so I did a run using KRPC and built my own autopilot features with it. It was a lot of fun, and I would recommend it.

So far I have done coding projects only, but I have some ideas for some non-coding agent projects in the future, so I'll give Gemma another shot

I did, but in almost everything I tried even qwen3.5 was better, and 3.6 was a huge step up.

I do love using local models when I can, but qwen-35B is the best model I can run, and while its an insanely good local model, it does not compare to the big ones.

I liked copilot because I didn't have to think about tokens. I get hung up when having to think about the price of things, and its hard to think about the project at the same time I got to think about token usage like a gas bill. The usage system had its own issues, but having a set amount of requests was a very comfortable way to use a paid AI service.

I haven't ran 27b that much because it only runs at like 2 tokens/sec on my computer.

I would say byteshape is smaller and faster, I can’t really notice a quality difference. But I haven’t used it as much as I only started using it a few days ago.

Now that I have tried out on a few tasks, Qwen3.6 is a huge jump in capability. It can make improvements to a project that qwen3.5 always struggled with.

I have been using Qwen3.5-35B-A3B a lot in local testing, and it is by far the most capable model that could fit on my machine. I think quantization technology has really upped its game around these models, and there were two quants that blew me away

Mudler APEX-I-Quality. then later I tried Byteshape Q3_K_S-3.40bpw

Both made claims that seemed too good to be true, but I couldn't find any traces of lobotomization doing long agent coding loops. with the byteshape quant I am up to 40+ t/s which is a speed that makes agents much more pleasant. On an rtx 3060 12GB and 32GB of system ram, I went from slamming all my available memory to having like 14GB to spare.

since they have been doing orchestral performances, getting to hear the orchestral versions of their old songs has been amazing.

I don't like the Netflix CGI slop movie style filter look it gives everything. But that is a more general trend in tv and movies that I just can't stand.

I do think this will eventually be a major part of the graphics pipeline, but I hope it will be limited and masked to things like hair, which is almost impossible to get right in real-time rendering.

without discussing copyright, I don't believe any of this is copied. Which I think should be the argument that actually matters.

I downloaded both 6.0 and 7.0 and based on only a light comparison of a few key files, nothing would suggest to me that 7.0 was copied from 6.0, especially for a 41x faster implementation. It is a lot more organized and readable in my armature opinion, and the code is about 1/10th the size.

Small local models are the only thing that still have that magic feeling to me. While large models are still useful and impressive, it makes more sense that they are happening on a giant super computer in a datacenter somewhere. But all the intelligence and capability that can run on my mid level gaming PC is astonishing to me.

Anonymous account unmasking represents a new threat to anonymity. not just this technique with llms, but the earlier text similarity one.

But I think it would be generally easier to counter in the same way.

Use an llm or heuristics to pose as someone else.

not only do you erase your traces, you add false positives in to the system which reduces the overall effectiveness of these techniques in the future. A bit of poisoning the well.

I hope eventually an easy to use tool, with maybe a small local llm, can make it easy enough to do this, so that any future deanonymization attacks would be too untrustworthy to rely on

I am curious on how you would algorithmically find the optimal solution for this kind of problem for much bigger grids. I wanted to do some seed finding in Factorio for the same exact problem using the generated map images, but never found a good solution that was fast enough.