The buy side takes the stage. Apply → https://t.co/f62TbpTQP2
Most blockchain conferences put builders in front of investors. This session is the reverse: two of the most senior allocator voices in digital assets on what institutions actually weigh, and what it takes to move.
Mark Cutis — Managing Director, Office of CEO Abu Dhabi Investment Council
Joe Marenda — Head of Hedge Fund Research and Head of Digital Assets Investing, Cambridge Associates
October 1, 2026 at the EYE Filmmuseum in Amsterdam. Limited spaces available, professional allocators only.
@Legends4Legends is organised by @ThetaInvest in support of Alternatives4Children, raising money for children's education.
This message is for information and education purposes only.
@MTorygreen The 30x gap makes sense. We use Claude Code daily and most commits just look like... commits. Nobody flags them as bot work because they go through normal review flows.
@arjunkhemani Financing is the easy part. Building and maintaining 9,000 km profitably is where it gets interesting. That maintenance budget alone would keep me up at night.
We use agents to perform increasingly complex tasks.
Now MirroS is exploring what happens when agents become evaluators too.
HarnessEval-W doesn’t simply apply the same metric to every world. It decides what needs testing, selects the right evaluation skills, gathers spatiotemporal evidence, and reasons over what actually happened.
The result isn’t just another number on a leaderboard.
You get an inspectable explanation of what failed and why.
That’s a much more useful direction for evaluating interactive worlds.
This photo was taken in Abu Dhabi… 💙
right in front of the Formula 1 circuit.
But the real race happening here has nothing to do with cars.
While the world is obsessed with building the next big AI model,
Abu Dhabi is playing a completely different game:
They’re building the infrastructure that will power the entire AI era.
→ $49 billion MGX AI fund
→ 1GW Stargate UAE cluster (part of a massive 5GW UAE-US AI campus)
→ National plan to put Agentic AI into 50% of federal government operations in just 2 years
Most people are still focused on algorithms.
Abu Dhabi is focused on compute, energy, data and scale.
That’s the real race.
And right now, they’re running ahead.
What do you think, are we underestimating infrastructure in the AI race?
#AI #AbuDhabi #Tech #Future #Stargate
What if years of complex computation could be reduced to hours or even seconds?
High-Performance Computing brings thousands of processors together to accelerate AI development, scientific simulations and data-intensive research.
For educational institutions, it means greater computing power without the cost and complexity of building heavy local infrastructure.
With secure, locally hosted supercomputing, Ankabut empowers UAE institutions to accelerate discovery, strengthen collaboration and contribute to national progress.
#Ankabut #HighPerformanceComputing #HPC #ArtificialIntelligence #ResearchInnovation #HigherEducation #UAETechnology
@Scobleizer Agree on expansion, but the J-curve is real. Uber's overhead ratios moved the wrong way despite 30% operating income growth. ROI might be there but lagging the spend.
This is one of the brightest application of AI I’ve seen. There’s massive talent in Nigeria and we won’t stop celebrating our young builders👏🏾👏🏾
RT this if it crosses your TL!
@LearnWithBishal Interesting demo but I'd want to run it on messier inputs. Real construction docs have contradictions, missing measurements, code issues. That's where these demos usually fall apart.
🤯 This AI wasn't supposed to be this good.
It's already outperforming ChatGPT, Claude, and DeepSeek in some benchmarks.
So I pushed it to the limit:
Build an entire Minecraft game with one prompt.
The result felt like a glimpse of where AI is heading next. 🧵👇
@Normal_2610 The centre debases, the state borrows, and the middle class pays either way. Capex cuts are just the slow-motion version of the same problem.
abu dhabi basically checks all three. cheap energy, 177 Mbps average, government actually wants data centers. bottleneck for me was banking KYC, not the build
Warren Hellman, co-founder of Hellman & Friedman
While there are almost no interviews, thankfully, there is a book where 26 pages describe his life before & the early days of building a $100B+ AUM private equity firm.
- All deals, including selling DoubleClick to Google and $2 billion in profit
- Throwing a free bluegrass festival for 700,000 people
- 3:45am morning run (every day!)
@MTorygreen Had this conversation with my team recently. We iterate on models faster than we can meaningfully test them. At this point the rate limiter isn't compute, it's how fast humans can read the output.
@hypersoren adverse selection cuts the other way here. Spirit's business failed, its operational data (schedules, pricing, maintenance logs) didn't. the company went bankrupt, the corpus didn't