Yesterday, just to prove to a friend that Jev isn't that "revolutionary" I extracted some image classification code that Claude had written for my private image organizer that used Qwen3-VL, and stopped the output at a single token, then used the probabilities. Input processing on my GPU was somewhere around 1000ms per image, so not too fast, but each question used the prompt cache, so followups were 100ms-ish.

That was my baseline of an untrained, non-optimized single pass classification. It would connect to my llama-server, and use the logprobs for the choices.

After that, I had Claude remove llama-server from the solution, and write it directly to the transformers, then I kept prompting it to profile and find more speed. Eventually my "Decision Engine" running locally on a trained 1B model got to just under 85% accuracy across my 500 validation prompts (images and text) not used or derived from the training set, and an 8MB image, with 10 questions with 5 choices per question, got down to just under 500ms. Pure text prompts and questions are below 100ms for 300tokens + 10 questions + 5 choices per question (average).

It did better on text than images, just because my training set included 90% text. I'll do more training and validation for images when I get home, but for now, I'm more than happy that I can get a local "decision engine" running in 4GB of VRAM and responding in under 30ms for most use cases I've had.

Even if the LLM thinks it’s made up D is the highest probability, that isn’t part of the set.

You never actually generate the prose, only the first pass, and grab the probabilities. It couldn’t ask for more details even if it wants to. It gets stopped before the first token renders.

I bought the Quest 3 as an upgrade because I had noticed my Quest 2 started having intermittent input lag issues. I never pinpointed what was causing it, but my Golf clubs would fly out of my hand and I'd have to wait a few seconds for them to eventually find their way back to me. My controllers would drift out of view when navigating menus. I uninstalled everything, reinstalled it again. When that didn't work, I upgraded.

The issues persisted in the new version as well. So for about the last 20-ish months, I've averaged under an hour a month, most months I don't even pick it up. I might do some homemade PC VR "experiments" (still have access to mouse and keyboard for the game, and not the controllers) on some random claude code experiments. I have a "cyberpunk" mobius strip space station walking sim I've been working on, that will eventually hook into a treadmill.

With that said, I've never actually used FoxPro. I only had a database as the contract for what my web app was supposed to do.

Everything costs more now than it did a year ago... except for THIS, and we are still complaining that a 90-99% reduction in cost is STILL too expensive. And a 50-75% reduction in time is STILL too long.

We used to have to wait for weeks for a design like that when I worked at a consultancy, and that is a week of salary. For the design, then it got handed off to a front end developer to slice it and get built so the back end developer can hook it up to a CRM. We are talking a month turn around with design, revisions, development, testing, and bug fixing.

It can now be done in a couple of hours for less than a single hour's cost. If it were 10x slower and 10x more expensive, it would STILL be "good deal".

From 2008-2017, “we” (I was a contractor 08-09, and employee 15-17) basically turned every spreadsheet into a web application internally.

We added authentication and authorization, used a LOT of ETL-type processing to move data around. So many things could have been packaged and sold, but that was the “secret sauce” that kept the company so profitable.

Then the end in 2017 when a new CIO came in, killed off IT (laid off all non-managers over the course of a year), and replaced everyone with South African consultants to turn IT from a cost center to a profit center.

He lasted another couple years then left.

She told me that "left to right" denoted progression in the story, "right to left" told the viewer the subject was "exiting" the current scene.

She didn't go into the details of WHY, and I probably didn't probe deeper, but it stuck with me, and I notice it all the time in film and television.

My job has been less watching Claude Code, and more watching Managers Claude Code.

I don't know which I hate more as a programmer.

I love space "browsers" - I lose myself in them, just clicking through all the entities.

Also TIL that there are satellite galaxies, and I've never noticed those on other toys/tools like this. I always though Andromeda was the closest galaxy, but it is FAR away compared to `Sgr dSph` which is within 50kly from the center of the Milky Way.

I'm sure most of it comes down to prompts, and all of them run over 100tps on a 3090. Smaller cards will likely be slower, but Qwen3.5 9B is small enough to fit on most consumer cards.

You can do each of those at various hosts and own nothing. Or own a couple "over priced" cards and do it all at home on battery power for a few hours while the power is out.

I'm not saying everyone has to run local LLMs, because the APIs are in a race to the bottom, and my $10 of OpenRouter credits I bought months ago is down to $8.94 because most models give you MILLIONS of tokens for a US Quarter.

I get that a lot of people don't have them. And a single one can be VERY performant. And the smaller models like a 7B can run on much smaller hardware like a mid-range [3|4|5]060.

My entire AI Dev Box cost $4500 in parts. 128GB RAM, i7-10700, 1TB and 2TB SSD, and 2x 3090s. Today's prices and inflation have definitely made that price tag seem a lot better than it was, but it was an investment in all things GPU that were happening in 2020 (crypto, blender, image gen), then LLMs exploded.

For code, they are great, but for creativity for NPC controllers, they leave something to be desired, but work well enough for testing, so I don't burn tokens until I'm actually playing my games.

But nothing one-shots a prototype better than Fable 5. I can have a prototype built in 30 minutes, hooked up to my local LLMs and Claude Code is very good at testing the interactions and even tuning the prompts of the NPCs for better experiences.

Even as a temporary consultant, at a company for only a couple of months, I remember seeing the CEO walk the hallway, greeting every single person on the way between a meeting room and the bathroom. I can barely remember my family member's names, or the last thing we talked about, let alone have the memory space for all that information to seem kind and courteous to "low level" employees.

In my entire career, I've never worked at a behemoth-sized company, though, so my experience has been much more "intimate" environments, and primarily in consulting firms or satellite offices away from the main HQ (even by only a couple blocks some times). So everyone in the office knew each other.

There are so many parts that are just "words" thrown together to sound technical. It reminds me of books you pick up in a science fiction game.