TechTonic Shifts - The AI Learning Hub | Podcast en RSS.com

TechTonic Shifts - The AI Learning Hub

TechTonic Shifts - The AI Learning Hub

por Marco van Hurne

Cursos Tecnología

Temporada 1

AI | ML news week 43

In this weekly podcast we discuss news and developments in the field of artificial intelligence. It covers new models from companies like Nvidia, Microsoft and OpenAI, as well as research papers focusing on areas such as video generation, web navigation and memory optimisation. We also talk about the rise of low-code AI platforms and the growing demand for developers who can tailor these platforms to specific industries and workflows.

AI | ML news week 39

TL;DR Another week, another batch of AI upgrades trying to take over the world (or at least your search results). From OpenAI’s fancy o1 models that "think" like chess grandmasters but move slower than molasses, to new open-source contenders like Alibaba’s Qwen 2.5 and Mistral’s Pixtral 12B. Microsoft even threw in a low-compute AI model, GRIN-MoE, which won’t make you mortgage your house for tokens, and much, much more

TechTonic Shifts' Brain Week

Brain Week looks at tech, brain science, and ethics. It covers many topics: Brain tech can collect and sell brain data. This includes sleep masks and meditation headsets. These products aren't as controlled as medical tools. This raises privacy worries. Brain-computer links let people control devices with thoughts. Neuralink and Synchron make these. They could help paralyzed people. But there are ethical questions. Some try to make computer chips work like brains. This could make AI faster and better. AGI means AI as smart as humans. It could bring benefits and risks. People debate if AI can be conscious. Views differ on this. Games affect the brain. They light up reward centers. Scientists use games to study thinking and choices.

AI enters the world of wearable tech

This podcast is all about how AI has entered the world of wearables. Think of the AI Compass or AI integrated within AR/VR glasses or in your Meta Quest, that sort of thing.

Agentic AI - a.k.a. AI Assistants

This is a podcast about Autonomous AI Agents This episode talks about how autonomous AI agents have evolved from their simple beginnings into self-thinking systems that could eventually run your business (or at least make your coffee). We’ll talk about: - The rise of autonomous agents — moving from just following commands to taking actions on their own. - Why tech leaders are getting on board - Practical use cases, from virtual assistants to robots that don’t just vacuum but also manage your calendar. The next step in AI is here, and it’s doing a lot more than just writing your emails. For more info on AI, check out https://techtonicshifts.blog

AI/ML news week 41

In this podcase, we talk about the major developments in the world of AI and ML. We discuss OpenAI's DevDay, where the company showed a lot of new developer features, including real-time speech-to-speech APIs, canvas interfaces for writing and coding, vision fine-tuning, prompt caching, and model distillation. We also discusses Nvidia's entry into the multimodal AI market with its NVLM 1.0 model, Meta's Movie Gen model for generating short videos from text prompts, and Microsoft's upgrade to its Copilot AI assistant. We conclude with a brief look at the research papers and articles that are shaping the future of AI.

AI/ML news week 40

In this week's AI and ML news, we discuss recent developments in artificial intelligence. We covers topics like new AI models from Google, OpenAI, and Meta, advancements in voice recognition and image generation, and the increasing cost of using AI tools.

The evolution of AI from Generation to Singularity

The podcast is based on a series of visionary articles written by Marco van Hurne, founder of the AI blog: TechTonicShifts.blog, discussing the evolution of artificial intelligence and its impact on business. In the podcast, we will talk about his framework that categorises AI development into four stages: Generation Space, Productivity Space, Business (re)Design Space, and the Singularity Space. In the podcast we talk about various AI applications in those spaces, including content creation, data analysis, trend spotting, autonomous communication, personal assistants, process automation, AI First businesses, Robotization, and ultimately Super Intelligence. It also examine the challenges and opportunities associated with these advancements.

When governments trust AI too much

Have you ever wondered what happens when governments put too much faith in AI? In this episode, we talk about the story of the Dutch Tax Department's "toeslagenaffaire" (social benefits scandal), which is a warning about the dangers of unchecked algorithms. A welfare program intended to support families transformed into a tool of injustice, which was tearing lives apart due to a rogue AI system. This isn't a dystopian future; it's the reality in the Netherlands The topics: How an algorithm, designed to detect fraud, became a weapon of mass destruction, wrongly accusing families based on flimsy factors like ethnicity and minor errors The tragic consequences: families losing their homes, children being placed in foster care, and even lives lost to suicide due to the crushing weight of false accusations and insurmountable debts The lone crusader, Pieter Omtzigt, who challenged the system and fought for justice, facing resistance and attempts to silence him The government's sluggish response in compensating victims, despite court orders to do so, exposing a blatant disregard for the law and a lack of accountability

Lessons how not to do AI by the Dutch Government

Up next, we've got a masterclass in how not to do AI, courtesy of the Dutch government. Because who needs AI ethics and oversight when you’ve got shiny new algorithms. We discuss two inglorious examples of bureaucratic brilliance that'll make you question if anyone in charge knows what AI actually stands for.

1 de 2