Looks awesome! It would be ever awesome-er if it had vision support too… 😀
LLMS
-
Karpathy Explains Claude Skills MCP Servers and AI Agents as New Baseline
By
–
Andrej Karpathy (@karpathy), OpenAI co-founder, ex-Tesla AI, "vibe coding" creator.
— Charly Wargnier (@DataChaz) 24 avril 2026
In just 4 mins, he explains why Claude Skills, MCP servers, and AI agents are past the hype and are now the new baseline for building.
Worth every second ↓ pic.twitter.com/Ef0oDov90gAndrej Karpathy (
@karpathy
), OpenAI co-founder, ex-Tesla AI, "vibe coding" creator. In just 4 mins, he explains why Claude Skills, MCP servers, and AI agents are past the hype and are now the new baseline for building. Worth every second ↓ -
Testing GPT 5.5 and New AI Technologies Before Review
By
–
And now I'm heading into the cave, because between GPT 5.5 and this I've got a ton to test out these days before bringing you a review
-
Model Pricing Strategy Compared to Private AI Competitors
By
–
The most attractive aspect of this model, and what puts the most pressure on it, is the price at which it is offered. Comparatively, this is how it would stack up against the private models it is compared to.
-
Token Budget Management: The New Organizational Challenge
By
–
Token budget as a management consideration is coming for every large org and almost none of them are ready to think about it the way they think about headcount or software licensing.
-
New Model 86% Cheaper Than Opus 4.6 With Identical Performance
By
–
the biggest thing is, it's 86% cheaper than Opus 4.6 at nearly identical performance.
-
DeepSeek V4 Pro: Advanced Open-Weights Language Model Released
By
–
Paper: https://
huggingface.co/deepseek-ai/De
epSeek-V4-Pro/blob/main/DeepSeek_V4.pdf
… Open Weights: -
DeepSeek V4: Largest Open-Source MoE Model Released
By
–
DeepSeek just dropped V4! Two open-source MoE models with 1M context windows under MIT license. DeepSeek-V4-Pro: 1.6T total parameters (49B active per token), pre-trained on 33T tokens. This makes it the largest open-source model available – bigger than Kimi K2.6 (1.1T) and
-
PRO Model Rivals Opus 4.6 and GPT 5.4 in AI Capabilities
By
–
The PRO model, as I say, rivals Opus 4.6, GPT 5.4, and Gemini 3.1 Pro in domains such as general reasoning and agentic capabilities. Private companies in the US cannot afford to stop.
-
DeepSeek PRO and FLASH: Large Models for Enterprise Value
By
–
DeepSeek lands in two sizes, a PRO and a FLASH. Overall, they are fairly large models, and I miss that family of smaller distilled models that made the famous R1 model saga so valuable. Even so, for companies this is gold!