concise summary of the shortcomings seen in existing AI models great use of examples to showcase how the under-performing properties manifest

100% of Jito Protocol Revenue Will Go To the DAO Jito just unveiled their most important product to date, and while there is lots to unpack from the product standpoint - an important market catalyst was revealed: "Jito Foundation and Jito Labs are collaborating on a JIP to direct *all fees* collected by Jito from BAM and the Jito Block Engine to the DAO Treasury." Currently, the DAO and Jito Labs split a collective 6% take rate on tips 50/50 (3% each), but this upcoming proposal would change that, directing all value towards the DAO. Using the current 30, 60, and 90 day moving averages, along with the median daily tip amount over the last year, the 6% take rate on Tips annualizes to the following: 30-D: $27,820,634 60-D: $25,876,009 90-D: $35,457,431 Median 1-Y: $39,912,695 In addition, the median weekly JitoSOL staking reward fee annualizes to an additional $9.3m for the DAO. Add in interceptor, a lesser known product, and the DAO is looking at approximately $45m-$50m in annual revenue for the DAO. Directing 100% of take rates on fees are set to bolster Jito's strongest existing revenue stream. On BAM: We will share some deeper thoughts within an upcoming research report, but our high level takeaway is that Jito's success has always come from obsessing over improving the user experience on Solana, and this is continued progress in that direction. They made a controversial decision in the past to deprecate their mempool, only to see that decision lead to a surge of activity on the network. And while moving away from an incumbent solution with a dominant market share is bold, we believe the continued desire to improve overall network functionality will, in hindsight, once again prove obvious for driving increased user activity. With BAM Jito is offering protocols an opportunity to enable more efficient, transparent, and decentralized block-building via plug-ins. We believe this will boost revenue and unlock a new wave of value creation for all stakeholders. In Summary, we several postiive tailwinds for Jito right now - Staked SOL ETFs - SubDAO on Token-Economics funded - 100% of Revenue going to the DAO (full alignment) - and now the Block Assembly Marketplace (BAM) We are excited to share more of our thoughts on all these topics in our upcoming research report. Be sure to follow and subscribe to our substack so you can get it first!

i vibe coded my childhood arcade dream in 5 hours. (to stop sitting 92hrs/wk)

who decided to start putting bananas into matcha lattes?

@artemis is one of the most underrated institutional-grade research imo link to their reports below

breakdown of V-JEPA 2 world model by @k7agar diving into the architecture that made it possible for it to carry out cup grasping with 65% success rate mentions about the 'language goal problem' where the robot is able to understand what it needs to achieve without being shown a pic / multiple pics would be interesting to explore a decentralized approach for that 1. world model generates iterations of 'goal' 2. decentralized verifier network votes on which one is regarded as an accurate 'goal' e.g. identifying BLT sandwich link below

Insightful @a16z episode on America’s energy problem Highlights • The need for employee retention for manufacturing and operating nuclear plants • Diving into the different energy types: solar and its strength, nuclear and the misunderstandings around it, wind and its downsides • The future of energy consumption, and the policies or advancements to help navigate the cyclical nature of energy demand (through the day, seasons) Episode ends with insights on future opportunities • AI software to manage the demand for energy • Decentralized coordination of energy resources A particularly interesting point was how weather could be a critical factor in the coordination systems Some projects building within the decentralized data collection space includes • @nubilanetwork • @SkyX_Network • @WeatherXM

@rayvo_xyz looking real sleek

1/N I’m excited to share that our latest @OpenAI experimental reasoning LLM has achieved a longstanding grand challenge in AI: gold medal-level performance on the world’s most prestigious math competition—the International Math Olympiad (IMO).

few learnings for builders: - if you’re building foundational models, you better move up the stack fast. - you must capture and leverage first-party data to drive personalization. - you must build connectors to further personalize with third-party data, but that won’t be differentiating for long — and will be harder after the impending bilateral data wars (see next edition of IMPLICATIONS) ;-) - you must capture some sort of social or professional graph to make recommendations/algos more relevant for the data you have. - if your system prompts are your moat, your days are numbered. - your context window must be increasingly wide and long, since “personalization effects” are the new network effects when it comes to retention and PLG. - markets will be won or lost at the interface layer, and the ultimate interfaces are ULTIMATELY controlled by the operating systems of our lives. for consumers, this is iOS and Android and our browsers (and maybe Meta and others to come?!). for enterprises, the OSes are our systems for work across every function - and those that value design and consider all the above learnings have such a great opportunity ahead…