INNOVATION SUMMIT 2026 @ESIC_University Excelente prospectiva de la evolución del contexto económico mundial por Ignacio de Gregorio @thewhiteboxAI. Estamos lejos de que la #IA sustituya el talento humano, lo potencia, pero cada vez será mas cara !!

At this point, the US market is just a leveraged play on whatever the fuck Taiwan and especially Korea indices do

I still haven't heard anyone explain to me how AI is going to go from barely scratching the first nine of reliability to the four or even five nines automation tasks will require to be economically viable. AI has cracked iterative workflows, but without automation, there's no trillion-dollar business.

As someone who has audited dozens of safety-critical systems, built static analysis tools, and used most formal verification and security tools, here are some red flags that should be a caution in taking these claims at face value: 1. There are no comparison benchmarks with 1/

Many are asking whether NVIDIA at $5 trillion is too much. But if Jensen's numbers yesterday are remotely true, he guided $400 billion in revenues for 2026 (without counting on gaming GPUs or non-operational revenues), which at NVIDIA's current net margin (will fall but alas), leaves $200 billion in net income, a forward PE of 25 for those $5 trillion, and at the current S&P500 average. Insanely as it sounds, if they deliver on the bookings number, it's not overvalued at all.

"We are being told that AI models are getting better, that AGI is close; yet we are getting worse and worse models for the same price. This tells us two things: we aren’t getting AGI soon, and that ‘fair’ is a charged word." - @TheTechOasis1 medium.com/p/the-great-ai-re…

This is exactly right. @TheTechOasis1

Why Most AI Products are Scams by @TheTechOasis1 medium.com/p/why-most-ai-pro…

The amount of cringe-posting from the “I had a few hours to be play around with GPT-5 and it <insert exaggeration>” AI influencer crowd that’s inundating my timeline tells me the release is probably waaaay less impressive than I would have hoped. It’s probably great, it’s also probably way too overblown.

What if you could not only watch a generated video, but explore it too? 🌐 Genie 3 is our groundbreaking world model that creates interactive, playable environments from a single text prompt. From photorealistic landscapes to fantasy realms, the possibilities are endless. 🧵

Energy-Based AIs, Too Good to be True? by @TheTechOasis1 medium.com/p/energy-based-ai…

AI, Education, & Jobs. We Need To Scrap Everything. thewhitebox.beehiiv.com/p/ai… via @TheTechOasis1

Meet Kimi-Researcher - an autonomous agent that excels at multi-turn search and reasoning. Powered by k 1.5 and trained with end-to-end agentic RL. Achieved 26.9% pass@1 on Humanity's Last Exam, 69% pass@1 on xbench. 🔗 Tech blog：moonshotai.github.io/Kimi-Re…

What If We Are All Wrong About AI? by @TheTechOasis1 medium.com/p/what-if-we-are-… #ArtificialIntelligence #Technology #DataScience #MachineLearning #Future

Apple’s Viral AI Paper. Reality or Fraud? by @TheTechOasis1 medium.com/p/apples-viral-ai…

Been heavily testing MCP with local and proprietary models, and I'm worried about memory complexity. As long as the cache has high memory complexity, preventing larger context windows, even the simplest of MCP use cases saturate the context window before I can do anything (this also applies to low-context providers such as, ironically, Anthropic). Either the MCP Servers are dumping too much context into my model, which is highly possible, or there's no way anyone can compete with Google's models; it's literally unusable technology if the context window isn't huge. If this doesn't change, I don't see how Transformer-based models can be of any use being run locally.

frontier ai today

MatFormers are very powerful alternatives to transformers. Similar to a regular transformer, but after training, you can split up the model to any size you like and get very strong performance that scales just like a regular transformer. So train once, get models of all sizes!

España tiene las segundas mayores reservas de uranio de la Unión Europea continental, como documenta el Libro Rojo del Uranio de la Agencia Internacional de la Energía. oecd-nea.org/upload/docs/app… Sin embargo, la extracción de minerales radiactivos como el uranio está prohibida por el artículo 10 de la Ley 7/2021, de 20 de mayo, de cambio climático y transición energética. Esta prohibición fue impulsada por el propio Gobierno, lo que le permite afirmar, de forma deshonesta, que «no tenemos uranio». boe.es/boe/dias/2021/05/21/p…

The Second ‘ChatGPT Moment’ is Here by @TheTechOasis1 medium.com/p/the-second-chat…