IronWolve

And claude isnt a replacement for the chain of humans, you dont know what you dont know, thats why we have a person in a role to gatekeep each step.

Anyone using claude can see it misses steps that an engineer wouldnt, it can make bad choices that a sysadmin wouldnt, it can pick the wrong order a project manager wouldnt, it can pick wrong cost models that a bean counter wouldnt, etc etc etc

But, I'm pretending it is and living with those mistakes.

I've known a few people that had their products copied by china and sold on amazon, ignoring patents. Same with copies of books. I've even bought a couple items just to find out they are knock offs. Its a big issue on amazon.

Think AI did this? Finding optimizations seems to be happening everywhere. Even llama.cpp is getting faster using AI, which they dont accept AI patches, so forks everywhere.

Anyone try this yet? reviews? thoughts?

When I first got my Amiga I spent all day one weekend copying PD game disks at the local computer store, Think it was like 25 cents to copy one plus price of a floppy. Good times.

It was fun to try out which local models made better agents. Few thoughts.

* locked to a single dir, so no system wide access.

* no local web search, can be fixed ddg or local mcp.

* no ssh, I want to have it ssh into my server and do the work.

* doesnt show the model being loaded, needs a bar/% counter.

* Can you drag/drop documents in the work dirs, or only + add them?

I love lm-studio, so cant wait to see how this goes. For local I normally use opencode + lmstudio.

Nice, I just made an IRC app that loads comic chat files a few weeks ago.

I added some updates, ignore bots, new expressions, diff message bubble behavior, more panels, custom avatars, colors, etc.

https://github.com/IronWolve/MaxChat

With AI and all the new projects/forks/updates, github is one busy place.

I wasn't browsing /trending on github until I started using AI, wanting to see what cool projects people are working on.

I noticed this, when it was only read a few files from my project, and I had to ask it to read ALL the files.

I then had it make a mistakes file and write every mistake, so it would learn, it kinda worked but it would still make the mistakes. It clearly wasn't reading all of it.

So I made a checklist, and it had verify every item on the checklist, that was my work around to both lazy and short mindedness of the agents. Turn mistakes into items to check for. Traded processing time for better results, ok for me on smaller projects. My run times went from 5-10 minutes from 3 per task, need to start logging tasks effectiveness/efficiency to reduce processing time.

I keep seeing people saying loop engineering is the way to get around these issues, I guess I'm kinda doing that in an adhoc way. Since I'm already looking at adding cost and goals(kinda).

I think things are moving fast, tested that new vibethink-3B, works on many small tasks/fast, and playing with ornith-35B with a draft vibethinker-3b as a draft gave me some good speed/results.

Was just trying to see how small I could go and get acceptable results, but yeah, larger Qwen 3.6 with MTP is going to be better. Cant wait to see how AI model (unsloth/local-llm/heretic/reaper/etc communities) are tweaking/engineering quality down into smaller models. Lots of new things coming out.

Somewhat true, the finances might fall apart, might, for some of the biggest companies, but AI is exploding in services industries of all sizes. Also, ai hardware is lasting way longer than 1 year... So a bubble? maybe, too many assumptions yet.

Having issues with coding a render for good looking realistic smoke coming off burning incense, opus 4.8 & gpt-5.5 both have code issues, glm-5.2 did it. Amazing.

The real time 3d fluid dynamics appear to be the tricky part, I wish I still had opus access, would love to see if it can do it.

Do people still use the K-Lite Codec Pack so their players have all the codecs installed? Or just use vlc?

I really liked when XP came out, and you could have both the clean 2000 look with advanced font rendering like cleartype. That was the perfect combo for desktop use.

I had my xp running a blackbox 4win and coLinux/cygwin, then moved onto vmware/virtualbox with windows vm and linux desktop, now I'm doing win11+wsl2 and loving it for cuda/ai/building projects.

I wonder. I spin up tons of projects, but I do it differently.

I'm also on windows. But I'm using a WSL linux vm to do the work. Then I have claude ./sync-to-windows, so the src is copied to a windows share via /mnt/drive/work and I just run a windows built.bat.

I have claude take care of the sync/build.bat so it can do things like clear cache, add options, look at logs, etc. This is also how I build multi platform, build on linux, src works on linux/osx/win

I'm a big fan of streaming radio, and been using apps for decades, so decided to make a modern version. Also, wanted to try AI coding out, and it worked out really well. Been having a blast.

https://github.com/IronWolve/StreamTuner-ng

I made my first vibe coded game, made casino style slots with lines, bets, free plays, auto play, in a linux shell, with color graphics, amazed how well it came out. Amazed how well it came out, with fixing graphics and adding features.

Might be nice on older games that dont have 4k HD texture packs, or even games like Skyrim or replaying older Witcher 1&2. Its upto the user right? Let people play older games with some better looking gfx. Seems like an easy win, use it or dont use it.

About 50-75%, also depends on cpu workload and brightness.

Running a VM will probably give you better life.

With all these long blog/news posts, been asking for a TLDR via AI. It does a good breakdown on thoughts vs actions. Then if it sounds interesting, then go back. Fabric app (oh github) works well ollama or lm-studio (and llm of choice)

It also works well with long YT videos using the closed captions, no time to listen to a 2 hour podcast.

TLDR:

* Trump and Elon Musk are reportedly using FCC Commissioner Brendan Carr to target speech they disagree with, posing a significant threat to First Amendment rights.

* Carr has been described as a tool in Trump's strategy to control free speech, with actions including investigations into major broadcasters like ABC, CBS, and NBC for their coverage or internal policies.

* The Verge's Nilay Patel criticizes this as an unprecedented attack on free speech, highlighting how the FCC, under Carr, might punish media for their editorial decisions, which contradicts traditional First Amendment protections.

* The narrative suggests that this administration's actions could fundamentally challenge the freedom of the press and speech in the U.S., using government authority in ways not seen before.

Whats the name of the books? Didnt see them in the article.

Pretty sure that latency and desktop improvement projects upped the timer for smoother mouse and windows movements also. Nice to see that it also improves application improvement of AI and encoders too.

I've been using LLM's to do searches for awhile, its quicker and i get better results. What happens in the future when new issues are only mentioned in github, x or reddit, and a different LLM is trained on each, have to use 3 searches?

Garmin Watch 30 days battery life

Fitbit 5-7 days

Apple 1 day

Grok has real time data, and its very good at data like mpg/price/etc. I use it as a search engine replacement, and it has good data.

so, it works for me

Again, asking a LLM to parse data isn't real life, thats not how life works.

And again, mentioning outsource/combine departs doesnt mean layoffs, and was not the objective of the query. If I ask it to give me the stats on gas crossovers with good mpg, I don't want a political lecture on why EV's are better for the environment.

LLM's shouldn't interject its own viewpoints when asked a question on data.

It was just a task to find issues and didn't even mention layoffs, it just assumed, just like you.

You do know asking chatgpt a question doesnt actually do anything in real life right?

From a legal viewpoint, selfcheck mistakes make you a criminal, and the stores make money for going after you for a settlement. Its already happening for customer mistakes to be prosecuted and settlements (walmart is the biggest culprit).

I'm buying groceries not getting into a possible legal issue, I'll avoid self checkout on a cart, maybe on 1 or 2 items self checkout is ok'ish.

Plus I feel using self checkout puts people out of jobs.

The political training in chatgpt has gotten in the way of asking funding and policy questions for basic questions.

I gave it a budget and asked it to give me the programs/departments/etc that have little return of value, overruns, possible fraud, spot problems, etc. so I could outsource or combine departments and save money.

It went on a long lecture that cutting funding was a horrible thing, and I'm horrible for asking, and refused to answer.

Really?

I'm asking basic auditing/restructure/spending, and it was trained to ignore my request and lecture against providing help, and refusal to give results.

smh, this isnt helpful.

I wonder if Apple runs statistics on controls and sees if a setting isn't used as much by the common user, so power users (audiophiles/etc), are hurt by these decisions.