@alexvoica.com
Corp affairs at @Synthesia.io, advisor at MBZUAI Guns N' Roses, basketball, Philip K. Dick, intelligent machines, peanut butter + maple syrup are a few of my favorite things. https://linktr.ee/alexvoica
What Europe is trying to do with AI regulation is, in principle, the right instinct.
But a lot of the rules on AI are written by smart academics, under the influence of lobbying groups trapped in a policy bubble that is heavily detached from the real world.
www.tiktok.com/@alexcvoica/...
6/ Thanks to the AWS policy team for inviting me to represent @synthesia.io, my fellow panellists for a lively conversation, and to everyone who grabbed me afterwards to chat!
5/ None of this is inevitable. If we want European AI champions and meaningful adoption, we have to make it easier to build here, scale here, and sell across borders. When half of your budget as an AI startup is spent on playing defense, it's hard to be effective on the offense.
4/ When the support system is missing, startups hesitate to build AI products and services, or they point their growth plans at markets that feel more innovation friendly.
3/ Second: is the EU’s digital single market delivering for an AI scaleup? Europe is excellent at writing complex rules from ivory towers. It’s less consistent at helping its own startups navigate those rules and win with them. (The AI Office has a habit of exhibiting this type of behavior.)
2/ Meanwhile the actual business runs on yesterday’s workflows. We need to stop treating AI like a tool, and start thinking about ways to redesign how work gets done, tied to outcomes people can measure and teams can own.
1/ First: why is Europe’s adoption of AI still broad but shallow? The short version is that too much of the middle layer is still parked in innovation labs, running experiments with chatbots and collecting screenshots to calm execs worried about AI adoption.
Just got back from the Amazon Expo in Brussels which focused on realizing Europe’s ambitions in AI and other technologies. On my panel, I tackled two questions that come up in almost every conversation about companies trying to “do AI" in Europe. 🧵👇
One of my favorite parts from the chat between @melissahei.bsky.social from the @financialtimes.com and @synthesia.io CEO Victor Riparbelli was this exchange about the future of AI video as a new format for learning and discovery: www.ft.com/content/2659...
From this day forward, please address me as "Head of corporate harmony at Synthesia."
Whenever someone who works at X or xAI posts on that cursed app, the first and only appropriate reply to anything they say is the Pedobear meme.
OpenAI: 10x engineers, 10x taste
Google DeepMind: 10x engineers, 1x taste
Anthropic: 10x engineers, 100x taste
xAI: 10x engineers, 0x taste
Meta Superintelligence Labs: 1x engineers, 0x taste
For the last few years we’ve lived in a weird split-screen reality: while AI labs have made wild breakthroughs in foundation models, productivity stats and corporate workflows have looked… basically the same.
Here's why that will change in 2026 👇
www.computerspeak.co/p/ai-predict...
Will OpenAI be the first company to build an AI consumer superapp?
Fidji Simo tried to build Facebook into a "traditional" superapp (e-commerce, dating, gaming, etc.) but it didn't work out.
Very interesting to see her try again, this time with better tech.
New: Grok is generating far more graphic sexual content than on X.
Grok’s image generator, Imagine, has been used to create explicit sexual videos of celebrities and violent sexual videos, a review of public posts shows.
70 Imagine URLs appearing to depict minors were sent to regulators today
We've just launched a landing page where everyone can create videos with our AI Santa for free: www.synthesia.io/santa
With a little help from our new AI video models, he’s now more realistic than ever, and ready to deliver personalized holiday cheer like never before!
With the AI Act, Brussels made a giant gamble: can you design comprehensive AI regulation before the technology settles down, and still stay competitive?
The answer is turning into a bitter lesson in what happens when process gets ahead of practice.
www.computerspeak.co/p/europes-bi...
Many of us have experienced that moment when we're creating a video, and just need the right image to illustrate a process or show something that supports our script.
Now you can create that cinematic visuals instantly in @synthesia.io with FLUX.2 - the new model from Black Forest Labs!
If you want to learn more about the impact of regulation on startups and other things that keep me awake at night when it comes to AI, I'll be talking about the State of European Tech report together at Slush later today (November 19): aws-experience.com/emea/north/e...
While a decade ago many startups treated comms as a service function that was run out of marketing to help with product growth and brand awareness, the work we do today is much more focused on risk mitigation and reputation management.
But that simple explanation hides another, more complicated reality: comms and policy are more important than 10 years ago because European elites have a distrust of the tech industry and are very hesitant to embrace emerging technologies.
For the first time in forever, Europe is producing globally relevant companies that are category leaders in what they do (or at least strong contenders): @synthesia.io, ElevenLabs, Lovable, Helsing, Black Forest Labs, Mistral AI - the list goes on.
Atomico’s annual State of European Tech report is out, and Anna Heim from @techcrunch.com noticed an uptick in scaleups playing a more active role in political debates: techcrunch.com/2025/11/18/e...
A 🧵 on why this is happening
European policymakers should pause and take the time now to engage directly with these entrepreneurs before implementation hardens into unintended disadvantage.
Synthesia is among them and our message is simple: don’t let unclear, overlapping and increasingly complex rules push the next wave of AI breakthroughs out of Europe.
Dozens of European founders, including many of whom are building generational AI companies that are global leaders in their fields, have already asked Brussels to stop the clock and listen.
A short, targeted rethink focused on clear, risk-based obligations and predictable guidance would do more for trust and innovation than pressing ahead with a patchwork that burdens smaller European builders.
The EU needs to rethink its AI rules so they actually strengthen, not sideline, European competitiveness and safety. There is still time to course-correct.
They were dead wrong. However, Europe’s political class became enamoured with these beliefs and got distracted by yesterday’s fears while forgetting about tomorrow’s opportunities.
