The EU is working to simplify its digital regulations, but its proposed Digital Fairness Act risks adding new complexity and uncertainty.
As @patrickgrady_ explains, rather than introducing more rules that could undermine innovation, EU policymakers should focus on consistent, evidence-based regulations that apply equally online and offline.
Truly bizarre exchange with Jacob
Coxon who says he will continue to enrich himself with AI lab stock despite his mass extinction fears. “I like money bro.”
Sick and tired of seeing charts in calendar time with no Y axis, extrapolated into huge findings. You need to look at the cost side, at the research inputs needed to maintain the progress. How does capability scale with research inputs, and does AI supply them? We see no strong sign that it does.
My post on this with @alexolegimasopen.substack.com/pub/silico…
This slide from the Anthropic IPO roadshow was a big factor in my decision to leave. It’s an insane plot and I’m surprised they made it public.
It shows the compute efficiency of pretraining over time. The y-axis is a **log scale**: they are seeing continued superexponential algorithmic progress.
Mythos was not a one-time deviation, but a point on a superexponential. They could already retrain Mythos with a fraction of the compute originally needed.
It doesn’t strictly imply a software-only intelligence explosion- they have been spending more on R+D. But algorithmic improvements do not require ridiculous creative genius and are on the verge of being automated. You do the math.
The European Commission has proposed a Digital Fairness Act that would address so-called “addictive design” on social media platforms.
As @patrickgrady_ warns, the approach conflates addiction with user engagement, while regulating platform design without clinical evidence risks creating uncertainty and inconsistent enforcement across Europe’s digital economy...
EU Kids Act Bulletin 3️⃣
In this edition: the EU pauses national social media initiatives, NGOs call for a tech-neutral approach to the KIDS ACT, and the EU’s parallel Digital Fitness Act is imminent.
eukidsact.substack.com/p/eu-…
Two years ago, Harari shrilled that the 2028 election would be run by AI. That prediction now looks even more ridiculous than it did back then, so he's back with another.
This time, alien invasion by 2031 🤷♂️
@TheEconomist – how about some sober voices?
boyswhocriedai.lovable.app/e…
Yuval Noah Harari: We are witnessing the end of human history. The end of human-dominated history. History will continue, but AI could be the one in control.
“In five years, there will be a technology that can make decisions and create new ideas independently.”
“Maybe they'll solve cancer and climate change. But we are not sure.”
“I'm tending to think of it more in terms of an alien invasion.”
“The aliens are not coming in spaceships. They're coming from the laboratory.”
“If humans are divided and in an arms race, it becomes almost impossible to contain this alien intelligence.”
Super Intelligence is just the latest makeover for 'AI.'
Like a wearisome Marvel franchise, we'd best prepare for the future rebrands of 'super' and 'ultra' AI.
halffullai.substack.com/p/th…
Europe has just 5% of global AI compute capacity, compared with roughly 75% in the United States.
Competing to build the largest closed AI models may be unrealistic. Open models offer Europe another path.
@patrickgrady_ explains ⤵️
vist.ly/5q7zt
EU Kids Act Bulletin 2️⃣
In this edition: confirmation that IMCO will lead the European Parliament's legislative work. Stakeholders welcome the proposal, but concerns remain over its breadth and impact on users' privacy.
Also—events to sign up to ✍️
eukidsact.substack.com/p/eu-…
Whether it's fixing security flaws, filtering spam, or taking down harmful content, AI continues to make the internet a safer place.
halffullai.substack.com/p/th…
Great new blog by my @ITIFdc colleague @patrickgrady_ on the EU Kids Act, arguing the regulation risks pushing children off the online communities they depend on without giving them viable alternatives.
It’s weird that the CEO of Hugging Face continues to be an “AI is normal technology” guy despite being the victim of an autonomous cyberattack by a swarm of rogue agents, arguably the least normal thing AI has done so far.