Marc Mojica

The Live Editor in #MATLAB supports a new plain text Live Code file format (.m) for live scripts as an alternative to the default binary Live Code file format (.mlx), but you can make (.m) as the default in the settings. Live scripts use a custom markup, based on markdown, where formatted text and the appendix that stores the data associated with the output and other controls. You can learn more here➡️ https://spr.ly/61103fAPpH

2

Designing a URL shortener is often dismissed as a "junior-level" exercise. But it remains a staple of the system design interview for one specific reason: it forces you to justify every single byte and millisecond of latency. It’s not just a coding challenge; it’s a high-stakes data management problem. At global scale, the "obvious" solutions quickly crumble under the weight of their own trade-offs. This series or articles on system design deep dives, focus not on the "how," but on the "why." In this article, we deconstruct:  • Why distributed counters are almost always superior to standard hashing.   • The real reason major shorteners choose less-efficient 302 redirects over 301s (it’s about product, not tech).   • The mathematical logic behind Base62 encoding.   • Caching strategies to protect your database from "viral tweet" traffic. If you are preparing for system design interviews or managing high-concurrency systems, check out the full breakdown below.

2

18 minutes here gets you a working Claude Code knowledge base built on nothing but markdown. Nate Herk implements Andrej Karpathy's LLM wiki idea step by step in Obsidian. Here's what it covers + timestamps: ✔ Karpathy's LLM wiki idea (01:40) ✔ Setting up Obsidian + Claude Code (05:39) ✔ Ingesting your first article (08:35) ✔ Querying and connecting projects (13:02) ✔ LLM wiki vs traditional RAG (15:36) Bookmark this for later, or watch it right now. Link in the comments

6

Stop training your AI on the "noise" of the open web. AI performance is only as good as the data behind it. Generic web scrapes often lead to hallucinations, weak domain expertise, and significant copyright risks, especially in high-stakes fields like medicine and engineering. AI Knowledge Nexus offers a trusted path forward. By leveraging 65B+ words of licensed, peer-reviewed scholarly content, Nexus provides the structured, machine-readable datasets your models need for high-performance training, fine-tuning, and RAG. Why AI Knowledge Nexus? - True authority: Full-text research from Wiley and premium partners (not scraped snippets ). - Deep domain expertise: Specialized coverage in science, healthcare, engineering, and finance. - Transparent attribution: Maintain clear, verifiable source attribution to support validation and compliance processes. Give your AI applications a foundation of knowledge you can actually trust. Ready to see it in action? 👉 Join our upcoming webinar to learn how scholarly content reduces risk and boosts model precision. 📅 January 29, at 3PM EST 🔗 https://ow.ly/Kzcl50Y36fg #Wiley #Nexus #AIWebinar #TrustedData #AITraining #AICompliance

15

1 Comment

The battle of agent frameworks continues. Microsoft just launched Microsoft Agent Framework. Use it to build, orchestrate, and deploy AI agents with support for both .NET and Python. Also supports multi-agent workflows with graph-based orchestration.

64

2 Comments

This has been my contention for a while, and I think the folks at Anthropic are in alignment. Chasing AGI is counterproductive. Specialize. Optimize for those specializations. There is an inflection point in any AI system where if you make it too large and too generalized, you will actually make the results worse over time. GraphRAG can help some there, but RAG involves creating chains that are effectively outside of the LLM meaning that they are still hampered if you don't have enough deep context. The problem with AGI is that it really has few benefits; human beings do not have true general intelligence; we tend to have clusters of specialised intelligences that overlap somewhat, which seems to be the case for both human/animal cognition and increasingly for LLMs.

35

4 Comments