1+ week, 1+ day ago   (321+ words) Databricks’ Proteus hit 1.8, 5.2x on Qwen 3.5 122B kernels by specializing for runtime shapes, but only after fixing reward-hacking in validation. Databricks says its Proteus harness generated Qwen 3.5 122B kernels that ran 1.8 to 5.2 times faster than the best vLLM implementations. Profiles of the companies…...

2+ week, 4+ day ago   (200+ words) Continue AI vs Wilco, compared side by side on 18 data points: what each one does, how big it is, what it has raised, the tech running each site, and what users report. We score both out of 100 on traction, team…...

2+ week, 5+ day ago   (216+ words) Ask-AI vs Glean, compared side by side on 21 data points: what each one does, how big it is, what it has raised, the tech running each site, and what users report. We score both out of 100 on traction, team and…...

3+ week, 11+ hour ago   (51+ words) Moonshot AI Funding Round · $700M raised · (2026) StartupHub.ai A mixed round — some signals are off the norm. Round size vs peers Far above the typical Funding Round round (58.3× the median) — an outlier. Heuristic read of round size vs peers, syndicate size,…...

3+ week, 1+ day ago   (325+ words) StartupHub.ai Alibaba's AI Spending Hits Profit, Meta-Microsoft AI Deal Bloomberg Tech covers Alibaba's profit plunge amid AI spending, Meta's deal with Microsoft, and Castellian's $13B valuation. Visual TL;DR. Alibaba AI Spending causes Profit Plunge. Alibaba AI Spending includes Alibaba…...

3+ week, 4+ day ago   (282+ words) Gabriel from Krea.ai discusses the infrastructure behind K2, detailing challenges in large-scale GPU training and innovative serving solutions. Visual TL;DR. Krea 2 Model requires In-house Training. In-house Training faced Training Challenges. Training Challenges addressed by Filesystem Strategy. Krea 2 Model served…...

3+ week, 4+ day ago   (686+ words) Fern (formerly Ishiki Labs) is a YC W2026 startup that pivoted from socially-aware meeting AI to custom robot world models for RL policy training. Here is how they did it, and how you can build a clone. Fern (formerly Ishiki Labs)…...

3+ week, 5+ day ago   (192+ words) The Qwen3 8-27B model has achieved impressive results on the Artificial Analysis Benchmark, demonstrating performance comparable to leading AI models like DeepSeek V4 and GPT-5.6 Luna Max. This development suggests that near-frontier AI capabilities are becoming more accessible. Qwen3 8-27B is a new artificial intelligence…...

3+ week, 6+ day ago   (81+ words) StartupHub.ai Founder & CEO at LlamaGen.AI Terry Zhang is a startup founder, storyboard artist, and builder based in SF/HK. He is the founder of LlamaGen.AI, an AI comic & manga platform, and previously co-founded AFFiNE. His work focuses…...

4+ week, 20+ hour ago   (51+ words) A solid round with healthy fundamentals. Round size vs peers Smaller than the typical Seed round. 2 investors on the cap table. All-new money — no prior investors re-upped. Mostly newer or unproven managers. Heuristic read of round size vs peers, syndicate…...