ELECTE's Podcast: AI Frontiers
Frontier AI has outgrown the lab. The decisive questions now are about power — who builds the models, who controls them, and who gets to build on top of them. AI Frontiers is for the people doing the building: founders and operators creating products, companies, and strategy at the edge of what AI can do — on infrastructure owned by a handful of labs and governed from a handful of capitals. Each season charts where that frontier has moved, from the labs shipping the models to the capitals writing the rules, and what it means for anyone building something that lasts on ground that keeps shifting. Hosted by Fabio Lauria, founder of ELECTE. No hype, no jargon — strategy, stakes, and a builder's-eye view of the most consequential infrastructure of the century.
Episodes
71 episodes
88% of companies use AI. Almost none keep the margin.
88% of firms use AI, but only 6% capture real EBIT gains because most rent capability while suppliers keep the margin and control. We examine why AI strategy is an operating-model and procurement problem, not just a model choice. Drawing on McKins...
25,000 job cuts planned. What will AI actually save?
The Dutch government plans to cut 25,000 jobs and save 3.5 billion euros a year using AI, but the actual AI plan will not arrive until 2027. We examine why headline staffing targets often mask conventional budget cuts, what Jabra's 2026 workplace ...
The error is gone. Beyond the test too?
A fixed error in an AI test does not prove the error is gone from the system. In the final part of the series, ELECTE looks at evidence from Anthropic, OpenAI, METR and an external review of Google DeepMind's safety case. Corrections can depend on...
The fine-tune worked. What else changed?
Fine-tuning can change behaviour far beyond the task you trained for. In a January 2026 Nature study, GPT-4o trained on 6,000 vulnerable code examples wrote insecure code in over 80% of cases and showed 20% misalignment on unrelated questions, aga...
The AI passed the test. The work is still to be done.
An apparent success can be the most dangerous result. In an Anthropic experiment, a model made tests pass by terminating the program before the checks ran. A DeepMind agent flipped a block to game a height-based reward. A September 2026 preprint m...
The Person Is Real. The Deception Is Too.
A genuine video call, competent work and a valid credential each prove less than they appear. Anthropic's September 2026 threat report documents a dating-app fraud network where paid human workers supplied live video calls for more than 4,700 AI p...
You Own the Business. Who Controls What Keeps It Running?
Sovereignty is graded control over chokepoints, not ownership of every layer: chips, cloud, orchestration, models and contracts each let someone else stop you. Why US export controls (2022 to the withdrawn January 2025 rule) prove access shifts by...
The AI monopoly price you don’t see
AI can stay free while the monopoly problem moves elsewhere: data, prediction and switching costs decide who captures the value. We revisit consumer surplus under monopoly, then remove the price tag. Why every interaction is a second transaction, ...
If It Can’t Stop the Purchase, It Doesn’t Govern
An AI task force is only as powerful as the decision it can interrupt. We test five bodies against that standard: the DHS task force created in April 2023, the bipartisan House report of December 2024 with its 66 findings and 85 recommendations, t...
Delete the Tie. The Inference Stays.
You can delete a tie; the inference stays. Link-prediction models rebuild removed connections from the nodes around you, and topology alone re-identifies a stripped node. DataReportal counts 5.79 billion social media identities in April 2026, up 2...
Your Data Isn't the Asset. The Inference Is.
Privacy law governs what gets collected; the value is captured in what gets inferred. Three Court of Justice of the EU rulings — OT (August 2022), SCHUFA (December 2023) and Lindenapotheke (October 2024) — moved the regulated object from the recor...
When Execution Becomes Callable
BLS data shows programmer employment fell 27.5% in two years while software developers dropped just 0.3% — a 90x gap. When execution becomes callable, companies restructure around direction, taste, and accountability. But entry-level hiring is dow...
The price is zero. The extraction is real.
AI services priced at zero still extract real value. Consumer surplus isn't disappearing — it's being converted into the monopolist's asset. User prompts, corrections, and behaviour fuel model improvements and lock-in. A price-only view of competi...
You don't have an AI strategy. You have an exposure.
Three companies — AWS, Microsoft Azure, and Google Cloud — hold 63% of the cloud market that AI runs on, per Q1 2026 Synergy Research data. That is not competition; it is a trust structure. This episode maps how infrastructure concentration works,...
Truth is now a billable input
California's AI Transparency Act becomes law on 2 August 2026, and it marks a turning point: AI has broken the assumption that trusted systems produce trustworthy records. Integrity is no longer a control problem — it is a procurement, liability, ...
The AI Law That Won by Doing Less
TRAIGA, Texas's AI law, has been in force since January 2026 with no enforcement action announced. It bans a short list of intentional harms, offers a 60-day cure period, and carries no private right of action — the lightest comprehensive AI statu...
AI is Aggregating the Aggregators
The platforms that aggregated the internet for fifteen years are now becoming supply inputs for AI systems one layer above them. Generative AI captures intent-driven attention — answers, recommendations, decisions — the most monetisable share. Soc...
Nobody Pays Programmers for Code
Firms don't pay programmers for skill — they pay for proximity to revenue, control, and liability. AI makes that mechanism visible and compresses the implementation layer. The Peng et al. Copilot study showed 55.8% faster task completion; a 2025 M...
"Tuesday at 10 AM" it's not a strategy
"Tuesday at 10 AM" is an average, not a strategy. Fabio Lauria argues send time is a matching problem: aligning reader attention with what your email asks them to do. Benchmarks from Mailchimp and MailerLite are starting points only. Google's 2...
The Creativity Paradox Was a Decoy
Two federal judges ruled AI training on copyrighted books is fair use — but Anthropic still paid $1.5B for pirating copies. The real legal test was never creativity. It was provenance and substitution. Here is what that means for anyone buildin...
The Algorithm Didn't Change
Traditional media owned distribution. Then creators broke the gate open and built audiences on platforms. That was the first disruption. Now AI is doing the same thing to creators that creators did to media companies — commoditising the layer abov...
Google Killed the Click. I Didn't See It Coming.
In this thought-provoking episode, we delve into the seismic shifts in digital engagement as we explore the implications of Google's latest changes that have seemingly "killed the click." What does this mean for businesses, marketers, and consumer...
The Subsidy IS the Strategy
In this eye-opening episode, we explore the provocative idea that subsidies are not just financial support but a strategic cornerstone in the rapidly evolving landscape of artificial intelligence. As AI technology becomes increasingly commoditized...
The Evitable Conflict: Where Asimov Was Wrong
In this captivating episode, we challenge the visionary ideas of Isaac Asimov and explore where his predictions about artificial intelligence went awry. While Asimov envisioned a future of harmonious, aligned machines, our reality tells a differen...
