Daniel de la Cuesta

Alan Kay, pioneer of object-oriented programming and the graphical user interface, talking about agents and tools in 1990.

Goose AI, the open source coding agent looks very interesting. Documentation is very well structured showcasing several useful use cases for this tool: goose.docs.ai You can build your own agentic infra for software development (coding agent + llm)

Gemma 4 just landed on the edge on Workers AI! 💎 MoE model with 26B and 4B active, for fast inference 💎 Tool calling, reasoning, vision capabilities. Generates code and is multilingual 💎 256k context window and Chat Completions compatible API 💎 Perfect for building fast agents developers.cloudflare.com/changelog/post…

All in, Ruby on Rails increases productivity for humans and agents

Obliquity, how complex goals are often achieved indirectly. Concept by John Kay youtube.com/watch?v=D9eobU…

For years I’ve shared predictions about how technology would reshape our world. We’re still early in that journey, but we’re getting closer every year. Here’s a look back at some of my predictions and the recent progress moving toward them:

An iOS QA agent is coming to a Cursor near you 🍎🤖👷 Announcing ios-simulator-mcp, a Model Context Protocol server for interacting with the iOS Simulator github.com/joshuayoes/ios…

One of the big shifts with large language models and Agentic AI is that code can now write code and actually get things done. In this new world, LLMs are basically the brain - the creative and planning part. But their creativity is still limited. They don’t grow the way humans do through experience or learning over time. Re‑training them is expensive, so their knowledge is basically read‑only once you get the model. Agents are like the body and the nervous system. They react to events, ask the LLM (the brain) what to do, and then carry out the plan using their own abilities or external tools. For now, there’s still nothing that gives AI real long-term memory or anything close to consciousness. The evolution of LLMs depends on new academic breakthroughs and deep research. Meanwhile, software engineers’ battlefield is with agents - the body. Our task is to build them smarter, faster, and more predictable using today’s state-of-the-art technology.

Will AI shift the traditional build vs. buy debate in favor of building? Low-code/no-code was a first step, but vibe coding could make that shift even clearer.

AI foundational models are beginning to look less like applications and more like operating systems. They are general-purpose tools you can program using plain English, not complicated commands or multiple clicks. You no longer need to install a specific app for every task; you just need to provide a description in words. This feels like the same type of inflection point we saw in the '90s when Windows made the PC a mainstream tool, or when smartphones disrupted the desktop. In the past, you would spend hours installing new software. Today, you can use the generic power of a foundational model to generate new functionalities on the fly.

As Andrew Grove described in "Only the Paranoid Survive" (goodreads.com/book/show/6686…), the IT industry has faced strategic inflection points in the past: from vertically integrated mainframes in the 80s-90s to horizontally specialized PCs (Intel chips, Windows OS, Dell assembly). After that came Internet and later smartphones with the iPhone and Android revolution. Today, generative AI and Large Language Models (LLMs) are driving another massive inflection point. Here’s how: - Chips & Microprocessors: The rise of specialized hardware like Graphic Procssing Unit (GPU) and Tensor Processing Units (TPU) is fueling AI’s computational demands, creating a new category of processors optimized for machine learning. - Operating Systems: Traditional OSes (Windows, Linux, iOS) are losing ground to foundational AI models (e.g., OpenAI’s GPT, Meta’s Llama, DeepSeek, Google's Gemini). These models act as new "operating systems," generalist and programmable via natural language. - Applications: Agentic systems and AI-powered apps like ChatGPT and CoPilot are disrupting SaaS and eating traditional software. Web building tools, to-do lists, business intelligence, spreadsheets, Enterprise Resource Planning (ERP) and even search engines face obsolescence as LLMs enable faster, more intuitive solutions.

@balajis They complement each other yes, Bitcoin is permission-less. AI agent could create a business and open a bitcoin wallet to store revenue.

In November 2022 the perception of crypto and AI couldn’t be more different, with FTX going down and ChatGPT booting up. “AI is the real deal, crypto is a scam” — that was the meme of the day. And I understand why. But of course, today AI is used for enormous amounts of slop. And crypto conversely is how we determine what’s actually real. …of course, both AI and crypto are very useful. I’m just discussing the shift in perception from Q4 2022 to now.

There are now two kinds of AI retards. The first kind of retard uses AI everywhere, even where it shouldn’t be used. The second kind of retard sees AI everywhere, even where it isn’t used.

How do you keep the agility and innovation of a small team while leveraging the resources and stability of a large corporation? What's your top tip? #Innovation #Leadership #Agile Interesting about Volkswagen software development company CARIAD: germanautopreneur.com/p/cariad-volks…

gpt-oss is a big deal; it is a state-of-the-art open-weights reasoning model, with strong real-world performance comparable to o4-mini, that you can run locally on your own computer (or phone with the smaller size). We believe this is the best and most usable open model in the world. We're excited to make this model, the result of billions of dollars of research, available to the world to get AI into the hands of the most people possible. We believe far more good than bad will come from it; for example, gpt-oss-120b performs about as well as o3 on challenging health issues. We have worked hard to mitigate the most serious safety issues, especially around biosecurity. gpt-oss models perform comparably to our frontier models on internal safety benchmarks. We believe in individual empowerment. Although we believe most people will want to use a convenient service like ChatGPT, people should be able to directly control and modify their own AI when they need to, and the privacy benefits are obvious. As part of this, we are quite hopeful that this release will enable new kinds of research and the creation of new kinds of products. We expect a meaningful uptick in the rate of innovation in our field, and for many more people to do important work than were able to before. OpenAI’s mission is to ensure AGI that benefits all of humanity. To that end, we are excited for the world to be building on an open AI stack created in the United States, based on democratic values, available for free to all and for wide benefit.

If you want to see growth and optimistic visions of the future you need to travel East.

Just returned from China (Beijing, Shanghai and Xian) - absolutely mind blowing! A society that proposes a compelling vision for a bright future based on big cities and technology.

Payments, super apps, electrification, transportation, infrastructure, energy, housing, …

Focusing their energies on solving big problems, no bullshit. Just trying to solve those big problems that affect many instead of small problems that affects only a few.