Brazilian alternative asset manager Patria and minority partner Ashmore agreed to sell one of Colombia’s largest solar parks to power generator Isagen for an undisclosed amount

BREAKING: HELL YEAH! CNN boldly defends their reporter Kristen Holmes after the White House brought up her kids to smear her after she dared ask Trump about his alleged mistress. This is how you stand up to this demented bully... “Kristen Holmes is one of the most respected and accomplished journalists covering the White House. This afternoon, she did her job and asked the President of the United States a tough, relevant, and newsworthy question on behalf of the American people," CNN said in an official statement. "Public officials are free to challenge reporting they disagree with, but personal attacks on journalists for asking questions are beneath the office and inconsistent with the principles of a free press. We stand firmly behind Kristen and reject these attacks in the strongest possible terms," the network added. Yesterday, Holmes asked Trump about a viral speech from Senator Jon Ossoff in which he stated that Trump "doesn’t want to do the job” and would rather “build his ballroom and travel with Natalie on their apparently defenseless flying palace gifted by the emir of Qatar." The reference to Harp — the "executive assistant" who many are now speculating might be Trump's mistress — enraged the President and he lashed out at Holmes, repeatedly telling her to be "quiet," calling her "disrespectful" and "fake news." The memo went out to MAGA land to back up Trump's attack, and the official White House Rapid Response account on X targeted Holmes, writing: "@KristenhCNN: Someday, your children will come across your disgusting and inhumane question. They will be sickened and embarrassed to have a parent be so callous and vindictive. It’s quite troubling." We commend CNN for standing up in the face of this disgusting smear campaign against an upstanding reporter. There is really only one big takeaway here: reporters need to start asking more questions about Natalie Harp. Please ❤️ and share to urge the press to ask more questions about Natalie!

I think some of the biggest uses of AI will come from work we currently don't bother doing. Too expensive. Not enough time. A potentially useful idea, but nobody available to investigate it. I recognize this from working in finance. You finish a reporting cycle knowing there are questions you would have liked to spend another day on. Then the next deadline arrives. You deal with what needs doing and leave the rest, even though a closer look might have been worthwhile. I wonder how that changes when investigating those questions costs very little. We would still need to check the work and take responsibility for the conclusions. But we could afford to ask more questions, follow up on smaller discrepancies and revisit assumptions that otherwise sit untouched for another year. The Jevons paradox argument is familiar by now. What interests me is what businesses actually choose to do with the lower cost. Reviewing every transaction instead of a sample, for example, or investigating a problem that was previously too small to justify someone's time. My bet is that there is a lot of demand we cannot measure yet because doing the work has never made financial sense. Reading the Anthropic Institute's paper on economic scenarios for AI brought me back to this. Its extreme scenario is extraordinary, but I found the middle one interesting enough: a substantially larger economy while most work still happens without AI. These are scenarios, of course. I wouldn't build a portfolio around a particular GDP number for 2030. What I take from it is that AI doesn't have to replace entire professions to have a very large economic effect. Then there is research, which is probably where my imagination runs furthest ahead of what I can put into a financial model. What could an engineering team attempt if it could explore many more designs before choosing which ones to build? How much could a small research team accomplish with analytical resources it could never previously afford? Physical testing would still take time and money. Plenty of ideas would fail. I just suspect we would attempt more, and eventually some of those attempts would produce things worth having. This is a big part of why I keep coming back to AI infrastructure. Broadcom and Marvell, memory, networking, power and cooling. We research these businesses separately, but they are serving different parts of the same system. If AI takes on more useful work across the economy, that system has to support it. The difficult part is figuring out who earns attractive returns from supplying it. Customers can get tremendous value from AI while an infrastructure shareholder earns very little. Our Flex analysis is a recent example of why I can like the business and still want a better entry price. But I remain extremely bullish on the broader opportunity. I suspect that, a few years from now, businesses will routinely do work they wouldn't even consider starting today. I'd rather judge the infrastructure opportunity against that possibility than assume today's chatbot habits tell us how much capacity we will eventually need.

Last Friday I had the pleasure of speaking at @BlockchainatCU! We covered what we do at @StrataMedia_, what life as a crypto researcher actually looks like, the biggest trends shaping the space right now, and how institutions are dipping their toes into the space. Below are a few tidbits and takeaways from the event 😁! Right off the bat, I want to thank everyone who came out! It was 7pm on a Friday, in the middle of midterm season, and we still filled seats! Great to see the passion and commitment from students to attend, especially on a topic as niche and controversial as crypto. It takes a driven individual to prioritize an event like this over the million +1 other things they could be doing. Not only that, but, during the talk, every time I looked up I saw engagement, interest, curiosity and a stark lack of cell phones. Really gives me faith that the future of crypto is in good hands. Throughout the night, we touched upon a number of topics, with the main points of interest being: > How do you approach research and create a strong thesis? > What makes a good crypto project in 2026 (hint: does the token accrue value?)? > And how far do we have left to go before we see peak institutional adoption (and what does that look like?)? I was also hit with a number of questions, primarily: > How do you evaluate good data vs bad data when multiple sources are giving different values? > Why is privacy so important in crypto and what is the tech behind it? > What is more valuable: the businesses being built today or the companies providing the "picks and shovels" behind their products? After the talk, the conversation continued for another half hour as students approached me to ask more questions, hear how I got into crypto, and learn how they can do the same. If I didn't have to leave I could easily have seen myself staying for another hour at least. For us in the industry, it's so easy to get caught up in the day to day grind: racing to meet deadlines, trying to keep up with the market, going to conferences and building "strategic" connections. But I remember not to long ago being one of those students, attending guest speaker events at my school's blockchain club, and hoping to get close enough to ask a few questions. I was so appreciative of the speakers who would show up, not because it would progress their careers, but because we were all students at one point and remember what it's like. And for the students out there, be curious and don't be afraid to reach out. I'd like to speak for the rest of the crypto professionals out there in saying that we love to hear from you. If you want to talk crypto, run through ideas, learn what working in the industry is really like, or need an introduction, my dms are always open. Lastly, for those of you reading this who were at the event, thank you. I feel so honored to have sat in front of you, and truly appreciate the eagerness, curiosity, and passion that you brought with you. Don't lose it!

Most wallet drains start with one signature people didn't understand. Paste it into Claude before you sign. "I'm about to sign this in my crypto wallet. Here is exactly what the wallet shows: [paste the message, contract address, function name and amounts]. I think I'm doing this: [for example, swapping 100 USDC for ETH]. 1. In plain English, what am I actually signing? 2. Does it let anyone move my tokens or NFTs? How much, and until when? Check for unlimited approvals, setApprovalForAll, and Permit or Permit2 signatures. 3. Does it match what I think I'm doing? 4. Any red flags: an approval much bigger than the trade, an unknown contract, a signature that can move funds without a transaction? 5. Verdict: looks normal, ask more questions, or don't sign. If you can't tell from what I pasted, say so." Already signed approvals in the past? Check and revoke them at It's open-source.

At Applied Compute most of our RL and eval pipelines grade with an LLM judge. These produce one result per rubric criterion, typically a paragraph of text accompanying a score. For some complicated tasks, grading can be a slow, pricey part of the loop. We swapped a grader on a code QA benchmark in AC2 for a decision grader: one call, a yes/no probability for every criterion at once. On 100 sample tasks (407 criteria), same answers: → ~32× cheaper
→ ~8× faster
→ 94% agreement with the LLM judge And it’s even better than the original judge: When the decision grader was confident, it matched the LLM judge 99% of the time. When it wasn't, the decision grader indicated low confidence (30-70%), pointing to ambiguous rubric items. This allowed us to identify noisy tasks and make the dataset cleaner. A cheap, fast, calibrated grader means you can grade more rollouts, ask more questions per call, and downweight or review the shaky calls instead of training on them.