I think many people still don't grasp what this means. Chris Miller analyzed in his book "Chip War" that China is about 10 years behind in lithography machine production. Meanwhile, Huawei chips are achieving roughly 80% of the performance of the NVIDIA H100. China is probably
@kimmonismus
-
OpenAI CFO Defends Growth as Anthropic Eyes Higher Valuation
By
–
OpenAI's CFO pushes back on reports of missed targets, claiming a "vertical wall of demand" , but admits the real bottleneck is compute, not product-market fit. Meanwhile, Anthropic might leapfrog OpenAI's $852B valuation, which says everything about how fast the moat narrative
-
Grok 4.3 Scores 53 on AI Index with Lower Pricing
By
–
Grok 4.3 is a very good model especially when you think its only 500m parameters! xAI's Grok 4.3 scores 53 on the Artificial Analysis Intelligence Index with ~40% lower input and ~60% lower output pricing vs Grok 4.20, making it one of the most cost-efficient models at its
-
Motif Neurotech Gets FDA Clearance for Skull-Top Brain Implant
By
–
Really exciting: Motif Neurotech just got FDA clearance to trial its DOT implant for treatment-resistant depression, and the key detail is where it sits: on top of the skull, not inside the brain. That alone could change the calculus for patients who need neurostimulation but
-
Prompt to Build Standalone Pac-Man Neon Arcade Game in HTML
By
–
/3 prompt: Create a single standalone HTML file for a complete playable Pac-Man-inspired neon arcade game. Output only final HTML code, with no explanations, no markdown, and no text outside the code. Use only HTML, CSS, JavaScript, and one full-page canvas. No external
-
Gemma 4 31B Beats Qwen 3 27B in Local LLM Gamedev Test
By
–
/1 Gemma 4 31B just crushed Qwen 3.6 27B in a local LLM gamedev contest inside @atomic_chat_hq (prompt is below)
— Chubby♨️ (@kimmonismus) 30 avril 2026
Device: MacBook Pro M5 Max, 64GB RAM
Results:
Qwen 3.6 27B: 32 tokens/sec · 18m 04s · 33,946 tokens
Gemma 4 31B: 27 tokens/sec · 3m 51s · 6,209 tokens
So what is… pic.twitter.com/wqyWyjXX2u/1 Gemma 4 31B just crushed Qwen 3.6 27B in a local LLM gamedev contest inside @atomic_chat_hq (prompt is below) Device: MacBook Pro M5 Max, 64GB RAM Results:
Qwen 3.6 27B: 32 tokens/sec · 18m 04s · 33,946 tokens
Gemma 4 31B: 27 tokens/sec · 3m 51s · 6,209 tokens So what is -
AI Video EXR Export Enables HDR Workflow in Nuke and After Effects
By
–
Resolution was never the problem with AI videos, it’s the dynamic range. Getting these as EXRs means the image stays clean while you're working in Nuke or After Effects, so you can finally use the results as real footage instead of a draft material. Run the SDR to HDR upgrade on
-
LTX-2.3 Reconstructs Blown-Out Dynamic Range in 8-Bit Footage
By
–
The sky in the original 8-bit footage here is completely blown out, but LTX-2.3 changed that, reconstructing compressed dynamic range.
— Chubby♨️ (@kimmonismus) 30 avril 2026
Not a filter. Not a tone-map. pic.twitter.com/qNz0NVnttLThe sky in the original 8-bit footage here is completely blown out, but LTX-2.3 changed that, reconstructing compressed dynamic range. Not a filter. Not a tone-map.
-
LTX-2.3 Upgrades SDR Footage to HDR with Scene Linear EXR Output
By
–
You can now upgrade existing SDR footage to HDR with LTX-2.3.
— Chubby♨️ (@kimmonismus) 30 avril 2026
Because it was trained on MP4 compression artifacts, it actually works on real world compressed video. It outputs scene linear EXR, giving you actual float data for DaVinci Resolve instead of a fragile image that… pic.twitter.com/TDqfSBCojBYou can now upgrade existing SDR footage to HDR with LTX-2.3. Because it was trained on MP4 compression artifacts, it actually works on real world compressed video. It outputs scene linear EXR, giving you actual float data for DaVinci Resolve instead of a fragile image that
-
Claude-Jupiter May Be Anthropic’s Next Sonnet-Class Model
By
–
Could "claude-jupiter" be Anthropic’s next Sonnet-class model? Sonnet 4.8 surfaced in the Claude Code source leak alongside Opus 4.7 and the Mythos/Capybara references — so this might be another internal codename rather than a random Discord rumor.