Redmond Hosting

🚨Hey everyone, I've got about 10 A100 servers available for rent.🚨 Each one comes with: 8× A100 80GB GPUs Dual 64-core AMD Milan 7763 CPUs 1TB RAM 8TB NVMe storage Uncapped internet Rent is $10,000 per month per server. If you're interested or want to know more, just send me a DM. #AI #MachineLearning #GPU #LimitedOffer

@lets_dig_deeper I work with an indian team, they speak Tamil. I am working on something that does live translations and I would love to test it out. We have a real issue to test against with native speakers, so I think we will be able to give great feedback.

need 5 devs to try out our silk api (alpha) and share feedbacks and co-work with the team comment if you want to be in our alpha test group

@janusch_patas LeechField? JK. Is it pronounced "Lick-Feld" for english speakers? Sorry if this is a dumb question.

Btw, it is called LichtFeld. Not LitchFeld, Litchfield, Lightfeld, LightField or any other variation or permutation.

@LLMJunky @ZypherHQ What’s the proper way to work from a markdown file?

@ZypherHQ Not as far as I'm aware , but I genuinely don't have this issue. Are you working from a markdown file? Or are you just letting it work from it's original context?

Look ma new Codex Updates! 0.119.0 and 0.120.0 are here. And with it, a HUGE number of quality of life updates and bug fixes! > Hooks now render in a dedicated live area above the composer. They only persist when they have output, so your terminal stays clean. If you're running PreToolUse or PostToolUse hooks, this is a huge readability win. > Hooks are now available again on Windows > CTRL+O copies the last agent output. Small but clutch when you're pulling a code block into another file or chat. > New statusline option: context usage as a graphical bar instead of a percentage. Easier to glance at mid-session when you're trying to gauge how much runway you have left. > Zellij support is here with no scrollback bugs. If you've been stuck on tmux just because Codex was broken in Zellij, you're free now (shout out @fcoury) > Memory extensions just landed. The consolidation agent can now discover plugin folders under memories_extensions/ and read their instructions.md to learn how to interpret new memory sources. Drop a folder in, give it guidance, and the agent picks it up automatically during summarization. No core code changes needed. This is the first real extension point for Codex's memory system, and it opens the door for third-party memory plugins. > Did you know, you can /rename a thread? But what's really cool about that is, after you rename it, you can resume it with the same name, no more UUIDs. codex resume mynewapp or directly from the TUI: /resume mynewapp > Multi agents v2 got an update to tool descriptions More reliable multi agent environments and inter agent communication > You can now enable TUI notifications whether Codex is in focus or not. Modify this in your config: [tui] notification_condition = "always" > MAJOR overhaul to Codex MCP functionality: 1. Codex Tool Search now works with custom MCP servers, so tools can be searched and deferred instead of all being exposed up front. 2. Custom MCP servers can now trigger elicitations, meaning they can stop and ask for user approval or input mid-flow. 3. MCP tool results now preserve richer metadata, which improves app/UI handoff behavior. 4. Codex can now read MCP resources directly, letting apps return resource URIs that the client can actually open. 5. File params for Codex Apps are smoother: local file paths can be uploaded and remapped automatically. 6. Plugin cache refresh and fallback sync behavior are more reliable, especially for custom and curated plugins. > Composer and chat behavior smoother overall, resize bugs remain though. > Realtime v2 got several significant improvements as well. > You're still reading? What a legend. 🫶 npm i -g @openai/codex to update

@MedOumoumad @Gertywood5 @ltx_model But when would a filmmaker want to remove a grade? They are the ones to drive the grades to begin with. I think this has a great use case for underexposed or blown out footage, but for home use, not for filmakers.

@Gertywood5 @ltx_model Exactly that yeah.

I Forgot to post about this one here "Ungrade" another IC LoRA I released a few days ago, this one shows just how @ltx_model with IC LoRA is re-defining what can be done or not in post-production. huggingface.co/oumoumad/LTX-2…

@SmorglesBord @Gaurab Also you need specific chasis for submersion. You can't just throw in servers meant for air cooling or normal water cooling. So its much bigger than just rack design. I think submersion is the future, I spent a year researching it. I wish it was talked about more.

@Gaurab There was a data center a few years ago that used submersion in mineral oil for cooling; of course that would affect rack design, which is probably why it is not used

It would take hurricane-force airflow above 90 decibels to air-cool a single NVIDIA GB200 NVL72 rack. The thermodynamic ceiling is 40 kW. That rack draws 120. Water absorbs 3,300 times more heat per unit volume than air. A V100 in 2017 drew 300 watts. An H100 in 2022 drew 700. A B200 in 2024 draws 1,200. GPU power quadrupled in seven years. Cooling infrastructure did not. CDU lead times went from 8 weeks to over 20. Vertiv invested $50 million in a single Ohio factory for liquid cooling. It will not be operational until 2027. CoolIT and Munters are sold out of high-capacity units. Microsoft, Google, Amazon, and Meta have power contracts signed, land secured, and GPUs on order. They cannot turn the racks on. The entire AI supply chain is waiting on pipe fittings.

@ChrisAtKIRI the output is clean, but at least on the huggingface demo the depth that it outputs is quite wrong.

Can you imagine how fast depth estimation has improved! InfiniDepth (github.com/zju3dv/InfiniD…) was accepted in CVPR 2026 and opensourced now. We used tsdf to create mesh directly to examine and you can see how clean the surfaces are. No refinement required at all. It is crazy. Congrats @pengsida #CVPR2026

@artificialguybr awe, you are too nice. I love what you do, if I come across some more GPUs you will always have a login :)

I would like to say a few words publicly. Closing a chapter that shaped me deeply. Two years, 150+ models released, and a partnership that changed everything , none of what ihave built could exist without @redmondai. Redmond journey taught me more than I could have learned anywhere else, and it helped me grow into the AI engineer and researcher I am today. When I was still launching my second model, doing my best with Google Colab, Graham came across one of my Reddit posts and believed in me. He never asked for anything in return. Grateful beyond words for the opportunity and the confidence he placed in me. Good luck to us both, and excited for what's next. 🙏

@kiwilikeklikek @CHENYAO981123 Battery powered?

@CHENYAO981123 looks good, do you have wireless or bluetooth version?

You guys can always count on me for the latest updates on MyActuator🤫🤫🤫

@AhmedShahnab so distorting a mesh based on brightness? its cool to play with stuff like this, but it is hardly a flex.

Introducing SYNESTHESIA: turning any video into real-time interactive 3D topographic light scapes How it works: For video uploaded, Every frame gets analyzed for luminance across 36,000+ points - brighter pixels rise into glowing peaks, with smooth lerping making the entire landscape breathe and morph Tech Stack: @Threejs + @reactthreefiber | Post-processing Music Credits: Dax - Dr. Dre ft. Snoop Dogg "Still D.R.E." Like if you’re impressed Repost if you’re into creative coding #CreativeCoding #ThreeJS #ReactThreeFiber #WebGL #GenerativeArt #DataVisualization #React #Frontend

@Humblepie5 @BarnettEwin @kimmonismus Considering the cost to install and maintain the rigid ones, I wonder what the ROI is on using rigid panels for the full 20 years, or just gluing on new flexible panels every 6 or 7 years.

@BarnettEwin @kimmonismus They don't last. Use the rigid type to get 20+ years of useful life.

A Chinese company has developed a flexible solar panel that can be directly adhered to rooftops, significantly simplifying installation. Solar energy is the future.

@gracia_vr @radiantimages I mean, this is more expensive than using regular cameras. You can absolutely get cameras that weigh the same or less than a phone, and you dont need to keep them all charged. Whats the real reason why you went with phones and not small and reliable cameras?

Exploring the potential of processing 4DGS from field captures rather than just studio setups, thanks to @radiantimages and their portable iPhone 17 Pro rig

@topazlabs @edmondyang Can you post closeups? Or links to HR videos with less compression? Its hard to see what astra can really do with these low-resolution, high-compression videos

The results our users are seeing with Seedance 2.0 + Astra are 🔥. Check out these 3 examples from @edmondyang:

@haider1 There wont be just one winner. Everyone is moving there together. Do you think if OAI gets to "AGI" first that means they won everything? No, a month later google will have an AGI that is slightly better, or faster.

AI bubble is popping openai is going to go bankrupt next year meanwhile, openai just raised a massive $110b round from Amazon, Nvidia, and Softbank that proves some people still don't get it. this race is worth funding even with the losses, because whoever wins AI ends up on top

@DougChampion @topazlabs @BillyBoman_AI Astra will work with any video source.

@topazlabs @BillyBoman_AI New media upscaled is wasted compute. The ONLY thing that matters is restorative passes of legacy media. Make sitcoms from the 90s look good, consumer OLED’s make SD look WORSE with every generation. NVIDIA is going to figure this out if you guys don’t. Why abandon this market?

You've seen Seedance 2.0, but NEVER like this. 🔥Seedance 2.0 + Astra for ultimate 4K quality. Check out these 9 examples from @BillyBoman_AI:

@MEYARachad lol It is kind of ridiculous that the plate rain is just falling straight down hahah.

5 days from scratch on this shot. Plate was just the car (+ a touch of rain). Everything else is full CG. 🎥👇 #VFX #Houdini #Solaris #USD #Karma #MadeByHumans

@_MaxBlade @michalkomar Yea its strange because it says I have a bunch of free credits, but it wont let me use any. I only used 3 saying "Hi" and now I cant test without paying.

@michalkomar Yess was scared people were gonna abuse tokens, should I give out more? and if so how much?

Introducing QuickClaw ✨ I made the first iOS app that launches your own OpenClaw agent in under 30 seconds. No Telegram No API keys No setup One app, Sign in → your agent is live. In the appstore right now. Limited capacity, first come first serve.

@_MaxBlade @1KTRDZ Can I put in my own api keys? That would add more weight to the free version

@1KTRDZ No, Your paying for the api credits code is free now. Compute is not my friend.

@drivascos It would be interesting if you could crop out the watermark and turn this into a giant splat.

the power of interactive camera controls to quickly visualize a scene 🎞️

@Jeremyseanwhite How does it compare to Embergen for speed?

HydroFX 1.0 is out and sets a new benchmark for GPU fluid simulation. I worked on the open ocean shark scene, nothing else comes close in speed, scale, or stability. Massive particles, white water passes, mesh export, fast iteration. Try the demo: storm-vfx.com/hydrofx

@aescripts Looks like Video Depth Anything.

Made with Depth Scanner by @_alanvfx aescripts.com/depth-scanner #aftereffects #aescripts #depthscanner Trying out depth scanner from @aescripts to get better volume and atmosphere entirely in comp. Pretty damn clean!