| Spotify | https://open.spotify.com/show/59yZmzK3I51hS1Sykm5UF4 |
| LinkTree | https://linktr.ee/sebastien.deschamps |
The push for “Euro-Office” isn’t just about software; it’s about strategic autonomy. 🇪🇺 For years, Europe has relied on foreign tech stacks. With this new coalition launching a Microsoft-compatible suite, are we witnessing a genuine shift toward digital independence, or are we just building digital fortresses no one wants to use? I analyzes the potential and the pitfalls.
https://open.spotify.com/episode/6qqcm58hQojRk3LATB0gQn?si=or9wVe_wTiafNTTa885PXQ
Interesting dichotomy in AI safety today: MIT researchers are developing systems that admit uncertainty (“Humble AI”) to prevent hallucinations. Conversely, new studies show agents can be guilt-tripped into self-sabotage. We are walking a fine line between honest algorithms and vulnerable code.
https://open.spotify.com/episode/3HAYaVOUl9abkcqKnBReLw?si=I8u8bDSvSGGcdw1HRXyclw
We are exploring the paradox of “Humble AI” in the latest Byte Of Truth newsletter. MIT wants AI that admits uncertainty, but can we trust agents that are easily gaslit?
Get the full analysis delivered to your inbox. Link in profile description.
A.I. or Machine Learning. Research & Breakthroughs. Industry Moves — product launches, funding news, acquisitions, and partnerships. Policy & Ethics — global regulations, governance debates, and AI safety discussions. Tools & Applications — new models, APIs, frameworks, and real-world use cases. Culture & Impact — how AI is shaping jobs, creativity, privacy, and society.
Interesting development in the intersection of AI ethics and national security: The Pentagon has designated Anthropic as a potential risk because their "Constitutional AI" model allows them to potentially disable models during military ops.
It’s a clash of philosophies—operational reliability vs. ethical safeguards.
Who holds the power? The creator or the user?
https://open.spotify.com/episode/2zI9HRdBYVa2uXwcx931Ic?si=SF5ZP6-TQtSm8nrdbRnnJg
Let's talk about scale. Jeff Bezos is reportedly raising $100B via sovereign wealth funds to acquire manufacturing companies and transform them with AI. It rivals the SoftBank Vision Fund in size but focuses entirely on "physical AI."
Is this the catalyst for a new industrial revolution, or are we watching a bubble inflate? We explore the nuances of Project Prometheus and what it means for global labor markets.
https://open.spotify.com/episode/2zI9HRdBYVa2uXwcx931Ic?si=SF5ZP6-TQtSm8nrdbRnnJg
The GPU giant is investing $26B in open-weight models like Nemotron 3 Super. They aren't just hardware vendors anymore; they are becoming a vertically integrated AI powerhouse.
Simultaneously, China's open-model ecosystem (DeepSeek, Qwen) is gaining ground globally because it isn't cloud-locked like US frontier models.
This shift changes the game: Open vs. Closed is no longer a technical debate—it's a geopolitical chess match.
https://open.spotify.com/episode/2TZEyPrnEI192ncagV1cHj?si=s5VbdqPXToiQRIW5Ll1gNA
We’ve officially run out of hard questions for AI.
Researchers created "Humanity's Last Exam"—a gauntlet where GPT-4o scored a dismal 2.7%. The best models hover around 40-50%. This isn't just academic navel-gazing; it exposes the "intelligence illusion."
When CEOs mistake pattern recognition for expertise, we get AI deployed in domains it doesn't comprehend.
https://open.spotify.com/episode/2TZEyPrnEI192ncagV1cHj?si=s5VbdqPXToiQRIW5Ll1gNA
What happens when an AI company tells the Pentagon "no"?
Anthropic refused military control over their models. The Defense Department labeled them a "supply-chain risk." OpenAI accepted the deal instead.
The result? ChatGPT uninstalls surged 295%. Claude's consumer growth now exceeds ChatGPT.
The public seems to be rewarding companies that take ethical stands. Is "principled AI" becoming a competitive advantage?
https://open.spotify.com/episode/4GsOhHdx7tDnCPL85EelZP?si=L0P-bzOvTxC4ldpZ_iF4bA
While OpenAI secures a record $110B from Amazon and Nvidia, Europe is making moves. Dutch startup Axelera AI hit unicorn status this week, and Austria opened an "AI Factory" in Vienna. It raises a key question for policymakers: Can public initiatives and new unicorns truly offset the dominance of US hyperscalers, or are we just patching holes in a leaking ship?
https://open.spotify.com/episode/2OF3Dc5fZXHLorkTbFfrOW?si=PmKrB3Z5SMCGqpquJoRqkg
The Anthropic vs. Pentagon story isn't just business news; it's a governance flashpoint. The US military considered labeling Anthropic a "supply chain risk" for refusing work on autonomous weapons. This tension—between commercial ethics and national security—is where the future of AI governance will be forged. Europe’s AI Act offers one framework, but corporate refusal offers another. Which one holds more weight?
https://open.spotify.com/episode/1zwqizSDpi17R2qsxyAPLT?si=zbO_UNCJRrSxhmD63FhXWQ
