From Dev.to Community
Tech Articles
Curated developer articles, tutorials, and guides — auto-updated hourly
Peremptory5d ago • 3 min read
Caltech's Physics AI Ditches Transformers for Neural Operators
Anandkumar and Jenik's new company uses operator learning instead of attention to handle 5 trillion ...
#research#modelrelease#aidevelopment#architecture
0 0
TechPulse 5d ago • 3 min read
Z.ai Unveils GLM-5.3, China’s New Frontier AI Model
Z.ai Unveils GLM-5.3, China’s New Frontier AI Model Tech news • Aug 14 2026 • 5 min...
#ai#zai#glm53#modelrelease
0 0
TechPulse 5d ago • 3 min read
Z.ai Unveils GLM‑5.3, the Fastest Frontier Model Yet
Z.ai Unveils GLM‑5.3, the Fastest Frontier Model Yet Meta description: Z.ai's GLM‑5.3 hits...
#ai#modelrelease#innovation#startup
0 0
AI Tech ConnectAug 23, 2026 • 1 min read
A 27B London Agent Beat Frontier Models at Reproducing Research
A 27B model, scaffolded well, reportedly beat far larger frontier models at rebuilding figures from ...
#modelrelease#research#ai#machinelearning
0 0
AI Tech Connect6d ago • 1 min read
The 27B Class: August's Small Open-Weight Models, Compared
Qwen3.8-27B, Nemotron 3.5 Lightning and Muse Glimmer 30B are the same size and almost nothing else. ...
#modelrelease#opensource#ai#machinelearning
0 0
AI Tech ConnectAug 22, 2026 • 1 min read
Gemini 3.7 Flash Arrives Three Weeks Later at Half the Price
Google-reported DeepSWE v1.1 rose from 49.0% to 65.3% in under a month, and the price halved. The in...
#product#modelrelease#ai#machinelearning
0 0
AI Tech Connect3d ago • 1 min read
Meta's Muse Code Bets on Cheap Tokens, Not Top Benchmark Scores
Meta's terminal coding agent loses every benchmark Meta itself published, and may still be the cheap...
#modelrelease#product#ai#machinelearning
0 0
Breach ProtocolAug 23, 2026 • 4 min read
Nobody can prove who built the stealth model everyone is testing
Ox Alpha, an anonymous model with a million-token context window that appeared in August, is still o...
#modelrelease#provenance#openweights#industry
0 0
AI Tech ConnectAug 24, 2026 • 1 min read
GLM-5.3 Found 1,097 Critical Bugs. Its Weights Are on Hold
Z.ai shipped GLM-5.3 without its open weights. The company says post-training produced a cyber capab...
#research#modelrelease#ai#machinelearning
0 0