renderartist

Switched from @claudeai Claude Code to @OpenAI Codex this week to see what all the hype was about. IMO Codex is woefully underwhelming and slow. I received a refund through an automated support chat and I'm headed back to Claude Code where I was able to get stuff done quickly. 😏

@RDKLInc Brave to take that on lol The oldest I was willing to go was a sunflower iMac, was able to snag one in great condition for $120. So fun revisiting OS 9 and Tiger on PPC, swapping the HDD to SSD, replacing PRAM batt and doing a general cleanup with new thermal paste this week. 👍🏼

An iMac Is Worth $1250...but Only With a Box!

@ostrisai Always appreciate seeing your experimental work, I don't think there is enough of it in our community. 👍🏼

I spent a few hundred dollars in compute on this experiment. Going to shut it down for now. I think it could potentially get there, but recovering the detail going from patch 2 to patch 4 would probably take much more time and money than it is worth, at least to me.

@emanueledpt Is Codex that much better than Claude Code? The back and forth is getting hard to keep up with, next month OAI will do something that triggers everyone and we'll be swapping it out for what, Claude Code again?

I'm cancelling my Claude 5x Max subscription. The models are ok, were better before. I just don't like paying the company who is so unclear about things Operates in the shadow Isn't clear in communication on what we can do or not with their sub When they asked for the feedback I said: "You fucked up too many times" I hope that they can wake up and do the right thing.

The amazing parts of Joshua Tree are on the trails, these generic shots are usually near the entrances and they’re all that most tourists bother to photograph. There are some super unique geological formations tucked back in the Wonderland of Rocks, Hidden Valley, and the Split Rock loop that most people will never see because they stopped at Skull Rock for a selfie and left. Another cool area of the park is the Lost Palms Oasis trail. When the ocotillo is in bloom it’s genuinely cool to witness. It’s a trek, but worth a visit at least once. Just bring A LOT of water.

Someone explain the allure of Joshua Tree to me like I'm 5. Just judging from the photos this would be a C- state park in Utah or Colorado.

I’m hearing there’s renewed lobbying in DC and in state legislatures to ban or severely restrict open-source. Like a few years ago, we’ll need everyone to help show policymakers why open-source matters: for startups, for competition, for economic growth, and for jobs. If you build with open-source, now is the time to speak up!

@gambala_codes DivX Angelfire Geocities Netzero Juno Demonoid KaZaa

Миллениалы, скажите что-нибудь на миллениальском, чтобы зумеры не догадались. Я начну: ВИНАМП.

Trying to get newbies up and running in ComfyUI is hard. We really need the Photoshop of AI models to interface in a truly unified way that eliminates guesswork and gets straight to function. Things are too clunky and it’s like re-inventing the wheel for every new model release.

@SrLinguicaVIVE hah love it.

O dev Lewilou22 criou um método para reproduzir filmes e programas de TV com seu Nintendo 64, usando flash carts. O processo gera um arquivo M1V para vídeo e um WAV64 para áudio, é necessária a conversão pois o N64 não suporta contêiner modernos. Tudo isso só foi possível graças libdragon, um kit de desenvolvimento de cód aberto para o Nintendo 64!

@ilyamiskov Such a clean clamshell unit, very cool. 👍🏼

Should I design a website/mobile app in Adobe Photoshop 5.0? 👀

I trained an ACEStep 1.5 XL LoRA on "some obscure 60s English rock band". Then I wrote a song about LoRA training and had them play it. Absolutely wonderful experience. I still have some UI work before I can make training public in AI Toolkit, but working on it as fast as I can.

@thekitze llama.cpp > vllm > ollama

wtf are ppl yapping about the ollama app is 39 times worse than lm studio in every single possible way? ui, model management, remote connections, chat, etc? what is this crap???

@aronprins @jdevalk 😂 All that junk will rear it's head in time, I'm sure. Just have Claude Code vibe code a custom SEO plugin for EmDash and adapt it as you go, skip like 80 telemetry calls and nags to upgrade just because you changed a description. Curious about e-commerce integration though.

@jdevalk So ehh… Yoast for EmDash? 🫣😜🤣🔥

I recently switched my blog to Astro away from WordPress. The comment I got most is "people still need a CMS". I didn't disagree, but did say this could all be easier with AI. Today, that's proven true, and I was lucky enough to have a sneak peek. #notaprilfools 1/3

Took a low-poly blockout and turned it into a high-res texture, optimized, game-ready mesh. Made entirely from Cursor in a few prompts, with the Scenario MCP (mcp.scenario.com) Now wrapping this into a one-click app on Scenario.

Download Toon-Tacular Qwen from @huggingface huggingface.co/renderartist/T…

Download Toon-Tacular Qwen from @HelloCivitai civitai.com/models/2499028…

Just published Toon-Tacular Qwen on @HelloCivitai and @huggingface - Check it out ⬇️

In these two images I'm using Ultimate SD Upscale in @ComfyUI with the full Qwen Image model + the lightning LoRA + my style LoRA to quickly upscale the final result while remaining close to trained style, it seems to work well as a refinement stage.

I'm intrigued by the models that have untapped quality potential, found Flux Klein to be too limited and same-y in that regard. IMO Qwen just seems to be more malleable in advanced workflows. It's the same reason I liked WAN 2.2 for image gen.

Tried a bunch of the latest image models and I'm still loving @Alibaba_Qwen Qwen Image 2512. In my testing there's just more faithfulness to the trained style – coherent and genuinely a joy to use.