The World Economic Forum has released a report on Asset Tokenization in Financial Markets.
Highlights
1. Tokenization offers a new model of digital asset ownership that enhances transparency, efficiency and accessibility.
2. This report analyses asset class use cases in issuance, securities financing and asset management, identifying factors that enable successful tokenization implementation.
3. Key differentiators include a shared system of record, flexible custody, programmability, fractional ownership and composability across asset types. These features can democratize access to financial markets and modernize infrastructure.
4. While the benefits are demonstrated, adoption is slowed by challenges such as legacy infrastructure, regulatory fragmentation, limited interoperability and liquidity issues.
5. Effective deployment requires phased approaches and strategic coordination among financial institutions, regulators and technology providers. Factors affecting design decisions – such as ledger type, settlement mechanisms and market operating hours – must also be carefully considered.
6. Ultimately, tokenization holds promise for a more inclusive and efficient financial system, provided stakeholders align on standards, safeguards and scalable solutions.
7. Tokenization is expected to reshape financial markets by increasing transparency, efficiency, speed, and inclusivity—paving the way for more resilient and accessible financial systems.
Highlights
1. Tokenization offers a new model of digital asset ownership that enhances transparency, efficiency and accessibility.
2. This report analyses asset class use cases in issuance, securities financing and asset management, identifying factors that enable successful tokenization implementation.
3. Key differentiators include a shared system of record, flexible custody, programmability, fractional ownership and composability across asset types. These features can democratize access to financial markets and modernize infrastructure.
4. While the benefits are demonstrated, adoption is slowed by challenges such as legacy infrastructure, regulatory fragmentation, limited interoperability and liquidity issues.
5. Effective deployment requires phased approaches and strategic coordination among financial institutions, regulators and technology providers. Factors affecting design decisions – such as ledger type, settlement mechanisms and market operating hours – must also be carefully considered.
6. Ultimately, tokenization holds promise for a more inclusive and efficient financial system, provided stakeholders align on standards, safeguards and scalable solutions.
7. Tokenization is expected to reshape financial markets by increasing transparency, efficiency, speed, and inclusivity—paving the way for more resilient and accessible financial systems.
❤4
Singapore's Sharpa unveiled SharpaWave, a lifelike robotic hand
—Features 22 DOF to balance for dexterity and strength
—Each fingertip has 1,000+ tactile sensing pixels and 5 mN pressure sensitivity
—AI models adapt the hand's grip and modulate force
—Features 22 DOF to balance for dexterity and strength
—Each fingertip has 1,000+ tactile sensing pixels and 5 mN pressure sensitivity
—AI models adapt the hand's grip and modulate force
HouseBots
Sharpa Unveils SharpaWave: The World’s Most Tactile Dexterous Robot Hand — HouseBots
Singapore-based robotics startup Sharpa is redefining what robotic manipulation means with the debut of its latest innovation: SharpaWave , a 22-degree-of-freedom (DOF) dexterous hand that brings human-like precision and speed to the world of robotics.
🔥6
Researchers introduced SPORT, a multimodal agent that explores tool usage without human annotation.
It leverages step-wise DPO to further enhance tool-use capabilities following SFT.
SPORT achieves improvements on the GTA and GAIA benchmarks.
It leverages step-wise DPO to further enhance tool-use capabilities following SFT.
SPORT achieves improvements on the GTA and GAIA benchmarks.
Google introduced Lyria RealTime is a new experimental interactive music generation model that allows anyone to interactively create, control and perform music in real time.
Available via the Gemini API and you can try the demo app on Google AI Studio.
Available via the Gemini API and you can try the demo app on Google AI Studio.
Amazon added AI-generated audio discussions about certain products, based on customer reviews and web searches.
Anthropic just now rolling out voice mode in beta on mobile.
Try starting a voice conversation and asking Claude to summarize your calendar or search your docs. Voice mode in beta is available in English and coming to all plans in the next few weeks.
Try starting a voice conversation and asking Claude to summarize your calendar or search your docs. Voice mode in beta is available in English and coming to all plans in the next few weeks.
Game-Changer for AI: Meet the Low-Latency-Llama Megakernel
Buckle up, because a new breakthrough in AI optimization just dropped, and it’s got even Andrej Karpathy buzzing)
The Low-Latency-Llama Megakernel a approach to running models like Llama-1B faster and smarter on GPUs.
What’s the Big Deal?
Instead of splitting a neural network’s forward pass into multiple CUDA kernels (with pesky synchronization delays), this megakernel runs everything in a single kernel. Think of it as swapping a clunky assembly line for a sleek, all-in-one super-machine!
Why It’s Awesome:
1. No Kernel Boundaries, No Delays. By eliminating kernel switches, the GPU works non-stop, slashing latency and boosting efficiency.
2. Memory Magic. Threads are split into “loaders” and “workers.” While loaders fetch future weights, workers crunch current data, using 16KiB memory pages to hide latency.
3. Fine-Grained Sync. Without kernel boundaries, custom synchronization was needed. This not only solves the issue but unlocks tricks like early attention head launches.
4. Open Source. The code is fully open, so you can stop “torturing” your models with slow kernel launches (as the devs humorously put it) and optimize your own pipelines!
Why It Matters ?
- Speed Boost. Faster inference means real-time AI applications (think chatbots or recommendation systems) with lower latency.
- Cost Savings. Optimized GPU usage reduces hardware demands, perfect for startups or budget-conscious teams.
- Flexibility. Open-source code lets developers tweak it for custom models or use cases.
Karpathy’s Take:
Andrej calls it “so so so cool,” praising the megakernel for enabling “optimal orchestration of compute and memory.” He argues that traditional sequential kernel approaches can’t match this efficiency.
Buckle up, because a new breakthrough in AI optimization just dropped, and it’s got even Andrej Karpathy buzzing)
The Low-Latency-Llama Megakernel a approach to running models like Llama-1B faster and smarter on GPUs.
What’s the Big Deal?
Instead of splitting a neural network’s forward pass into multiple CUDA kernels (with pesky synchronization delays), this megakernel runs everything in a single kernel. Think of it as swapping a clunky assembly line for a sleek, all-in-one super-machine!
Why It’s Awesome:
1. No Kernel Boundaries, No Delays. By eliminating kernel switches, the GPU works non-stop, slashing latency and boosting efficiency.
2. Memory Magic. Threads are split into “loaders” and “workers.” While loaders fetch future weights, workers crunch current data, using 16KiB memory pages to hide latency.
3. Fine-Grained Sync. Without kernel boundaries, custom synchronization was needed. This not only solves the issue but unlocks tricks like early attention head launches.
4. Open Source. The code is fully open, so you can stop “torturing” your models with slow kernel launches (as the devs humorously put it) and optimize your own pipelines!
Why It Matters ?
- Speed Boost. Faster inference means real-time AI applications (think chatbots or recommendation systems) with lower latency.
- Cost Savings. Optimized GPU usage reduces hardware demands, perfect for startups or budget-conscious teams.
- Flexibility. Open-source code lets developers tweak it for custom models or use cases.
Karpathy’s Take:
Andrej calls it “so so so cool,” praising the megakernel for enabling “optimal orchestration of compute and memory.” He argues that traditional sequential kernel approaches can’t match this efficiency.
hazyresearch.stanford.edu
Look Ma, No Bubbles! Designing a Low-Latency Megakernel for Llama-1B
🆒5
Telegram + Grok = this summer https://xn--r1a.website/durov/422
Telegram
Pavel Durov
🔥 This summer, Telegram users will gain access to the best AI technology on the market. Elon Musk and I have agreed to a 1-year partnership to bring xAI’s chatbot Grok to our billion+ users and integrate it across all Telegram apps 🤝
💪 This also strengthens…
💪 This also strengthens…
🆒5
Apple and Duke University introduced 𝐈𝐧𝐭𝐞𝐫𝐥𝐞𝐚𝐯𝐞𝐝 𝐑𝐞𝐚𝐬𝐨𝐧𝐢𝐧𝐠
Researchers train LLMs to alternate between thinking & answering.
Reducing Time-to-First-Token (TTFT) by over 80% AND improving Pass@1 accuracy up to 19.3%!
Researchers train LLMs to alternate between thinking & answering.
Reducing Time-to-First-Token (TTFT) by over 80% AND improving Pass@1 accuracy up to 19.3%!
Market map for browser agents
A new companies launch in the space every week, for both consumer and enterprise use cases. ManusAI is one of the most popular generalist consumer agents, and Athena Intelligence is already being used by companies like Anheuser-Busch.
Computer/browser use has become one of the most important frontiers for model capabilities, with OpenAI, Anthropic, and Google DeepMind having dedicated teams to Operator, Claude Computer Use, and Project Mariner.
Open source frameworks like Browser use and Stagehand have become some of the most popular repos on Github, with tens of thousands of stars.
AI-first browsers are poised to disrupt the massive web browser market, with highly anticipated releases like Comet from Perplexity on the way. It's yet to be seen how Google integrates Project Mariner and other AI tools within Chrome.
A new companies launch in the space every week, for both consumer and enterprise use cases. ManusAI is one of the most popular generalist consumer agents, and Athena Intelligence is already being used by companies like Anheuser-Busch.
Computer/browser use has become one of the most important frontiers for model capabilities, with OpenAI, Anthropic, and Google DeepMind having dedicated teams to Operator, Claude Computer Use, and Project Mariner.
Open source frameworks like Browser use and Stagehand have become some of the most popular repos on Github, with tens of thousands of stars.
AI-first browsers are poised to disrupt the massive web browser market, with highly anticipated releases like Comet from Perplexity on the way. It's yet to be seen how Google integrates Project Mariner and other AI tools within Chrome.
🔥4
New paper from Google DeepMind Beyond Markovian: Reflective Exploration via Bayes-Adaptive RL for LLM Reasoning
Researchers study 𝙬𝙝𝙮, 𝙝𝙤𝙬, and 𝙬𝙝𝙚𝙣 LLMs should self-reflect and explore at test time—questions that conventional Markovian RL cannot fully answer.
HuggingFace
GitHub
Researchers study 𝙬𝙝𝙮, 𝙝𝙤𝙬, and 𝙬𝙝𝙚𝙣 LLMs should self-reflect and explore at test time—questions that conventional Markovian RL cannot fully answer.
HuggingFace
GitHub
🆒4
An open-source humanoid for under $3k. Meet HopeJr, a full humanoid robot lowering the barrier to entry
Capable of walking, manipulating many objects, open-source and costs under $3000.
Designed by Rob Knight and HuggingFace.
Full bill of materials and links to source the parts will be available on this github
HopeJr has 66 actuated degrees of freedom.
Capable of walking, manipulating many objects, open-source and costs under $3000.
Designed by Rob Knight and HuggingFace.
Full bill of materials and links to source the parts will be available on this github
HopeJr has 66 actuated degrees of freedom.
GitHub
GitHub - TheRobotStudio/HOPEJr: HOPEJr_open-source_DIY_Humanoid_Robot_with_dexterous_hands
HOPEJr_open-source_DIY_Humanoid_Robot_with_dexterous_hands - TheRobotStudio/HOPEJr
#DeepSeek-R1-0528 is here
- Improved benchmark performance
- Enhanced front-end capabilities
- Reduced hallucinations
- Supports JSON output & function calling.
Weights
- Improved benchmark performance
- Enhanced front-end capabilities
- Reduced hallucinations
- Supports JSON output & function calling.
Weights
Deepseek
Your First API Call | DeepSeek API Docs
The DeepSeek API uses an API format compatible with OpenAI/Anthropic. By modifying the configuration, you can use the OpenAI/Anthropic SDK or softwares compatible with the OpenAI/Anthropic API to access the DeepSeek API.
SEO is slowly losing its dominance. Welcome to GEO.
The future of search, marketing, and performance in the LLM era. As search changes, a new paradigm is emerging in marketing, one driven not by page rank, but by language models. Enter Generative Engine Optimization (GEO).
In the age of ChatGPT, Perplexity, and Claude, GEO is positioned to become the new playbook for brand visibility. GEO is rewriting the rules of search — unlocking an $80B+ opportunity.
It's not about gaming the algorithm — it's about being cited by it.
The brands that win in GEO won't just appear in AI responses. They'll shape them.
The future of search, marketing, and performance in the LLM era. As search changes, a new paradigm is emerging in marketing, one driven not by page rank, but by language models. Enter Generative Engine Optimization (GEO).
In the age of ChatGPT, Perplexity, and Claude, GEO is positioned to become the new playbook for brand visibility. GEO is rewriting the rules of search — unlocking an $80B+ opportunity.
It's not about gaming the algorithm — it's about being cited by it.
The brands that win in GEO won't just appear in AI responses. They'll shape them.
Anthropic just open-sourced their AI Mind-Reading Tools
Anthropic has just released their groundbreaking circuit tracing tools to the public, and this could be a game-changer for understanding how large language models actually "think."
What's the Big Deal?
For too long, AI models have been black boxes—we know what goes in and what comes out, but the internal decision-making process remains mysterious. Anthropic's new tools change that by creating attribution graphs that reveal the step-by-step internal processes models use to generate their outputs.
Think of it as getting a peek inside the AI's "thought process" as it works through a problem.
What You Can Do Now
The open-source release includes:
1. Circuit Tracing Library - Generate attribution graphs on popular open-weights models 2. Interactive Frontend - Explore graphs visually through Neuronpedia's interface 3. Hypothesis Testing - Modify feature values and see how outputs change in real-time.
As Anthropic CEO Dario Amodei recently emphasized, our understanding of AI internals is lagging dangerously behind AI capabilities. This transparency gap is a real problem as AI systems become more powerful and prevalent.
By democratizing these research tools, Anthropic is enabling the broader community to:
- Study multi-step reasoning processes
- Understand multilingual representations
- Discover new behavioral patterns in AI models
- Build safer, more interpretable AI systems.
The team has used these tools to analyze interesting behaviors in Gemma-2-2b and Llama-3.2-1b models, with demo notebooks showing real examples of circuit analysis in action.
Researchers can use the Neuronpedia interactive interface here.
GitHub.
Anthropic has just released their groundbreaking circuit tracing tools to the public, and this could be a game-changer for understanding how large language models actually "think."
What's the Big Deal?
For too long, AI models have been black boxes—we know what goes in and what comes out, but the internal decision-making process remains mysterious. Anthropic's new tools change that by creating attribution graphs that reveal the step-by-step internal processes models use to generate their outputs.
Think of it as getting a peek inside the AI's "thought process" as it works through a problem.
What You Can Do Now
The open-source release includes:
1. Circuit Tracing Library - Generate attribution graphs on popular open-weights models 2. Interactive Frontend - Explore graphs visually through Neuronpedia's interface 3. Hypothesis Testing - Modify feature values and see how outputs change in real-time.
As Anthropic CEO Dario Amodei recently emphasized, our understanding of AI internals is lagging dangerously behind AI capabilities. This transparency gap is a real problem as AI systems become more powerful and prevalent.
By democratizing these research tools, Anthropic is enabling the broader community to:
- Study multi-step reasoning processes
- Understand multilingual representations
- Discover new behavioral patterns in AI models
- Build safer, more interpretable AI systems.
The team has used these tools to analyze interesting behaviors in Gemma-2-2b and Llama-3.2-1b models, with demo notebooks showing real examples of circuit analysis in action.
Researchers can use the Neuronpedia interactive interface here.
GitHub.
Anthropic
Open-sourcing circuit tracing tools
Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems.
❤8🔥2👏2
Sakana AI introduced the Darwin Gödel Machine: Open-Ended Evolution of Self-Improving Agents
Researchers harness the power of open-ended algorithms to search for agentic systems that get better at coding, including improving their own code.
It’s the Automated Design of Agentic Systems (ADAS), but where it also edits itself. Experiments show both the ability to self-improve and the open-ended search are essential. If done safely, DGMs could accelerate AI development and allow to reap its benefits much sooner.
Report.
Code.
Researchers harness the power of open-ended algorithms to search for agentic systems that get better at coding, including improving their own code.
It’s the Automated Design of Agentic Systems (ADAS), but where it also edits itself. Experiments show both the ability to self-improve and the open-ended search are essential. If done safely, DGMs could accelerate AI development and allow to reap its benefits much sooner.
Report.
Code.
sakana.ai
Sakana AI
The Darwin Gödel Machine: AI that improves itself by rewriting its own code
🆒5
Stanford introduced the first general-purpose biomedical AI agent
Biomni is a free web platform where biomedical scientists can immediately delegate their tasks to Biomni.
Biomni automates literature reviews, hypothesis generation, protocol design, bioinformatics analysis, clinical reasoning, and much more — scaling biomedical expertise for 100× the number of discoveries.
Key results:
1. Designed a cloning experiment with real-world wet-lab validation; on par with 5+ year expert in a blind test
2. Ran 458-file wearable bioinformatics analysis in 35 minutes vs. 3 weeks (800x faster) for human expert
3. Uncovered novel hypothesis: new TFs regulating skeletal lineages on a large scRNA+scATAC data
4. Human-level performance on LAB-bench DbQA and SeqQA, with SOTA at Humanity’s Last Exam and across 8 new biomedical tasks
5. Biomni-E1 – the first unified environment designed for a biomedical agent—encompassing 150 tools, 59 databases, 106 software—systematically curated from 2,500+ bioRxiv papers
6. Biomni-A1 – a generalist agent with retrieval, planning, and code as action
Paper
Code.
Biomni is a free web platform where biomedical scientists can immediately delegate their tasks to Biomni.
Biomni automates literature reviews, hypothesis generation, protocol design, bioinformatics analysis, clinical reasoning, and much more — scaling biomedical expertise for 100× the number of discoveries.
Key results:
1. Designed a cloning experiment with real-world wet-lab validation; on par with 5+ year expert in a blind test
2. Ran 458-file wearable bioinformatics analysis in 35 minutes vs. 3 weeks (800x faster) for human expert
3. Uncovered novel hypothesis: new TFs regulating skeletal lineages on a large scRNA+scATAC data
4. Human-level performance on LAB-bench DbQA and SeqQA, with SOTA at Humanity’s Last Exam and across 8 new biomedical tasks
5. Biomni-E1 – the first unified environment designed for a biomedical agent—encompassing 150 tools, 59 databases, 106 software—systematically curated from 2,500+ bioRxiv papers
6. Biomni-A1 – a generalist agent with retrieval, planning, and code as action
Paper
Code.
biomni.stanford.edu
Biomni - A General-Purpose Biomedical AI Agent
A general-purpose biomedical AI agent to automate biomedical research.
❤4🔥2
Stablecoin infrastructure reaches enterprise adoption phase
Fireblocks' latest survey of 295 financial executives reveals that stablecoin adoption has moved beyond experimentation into operational deployment, with 90% of respondents actively implementing or planning stablecoin payment systems.
Key Market Dynamics
The research shows a fundamental shift in priorities. Speed has emerged as the primary value proposition, cited by 48% of respondents, while cost savings ranked lowest at 30%. This suggests organizations view stablecoins as performance enhancers rather than cost-cutting tools.
Infrastructure readiness has reached a tipping point, with 86% of firms reporting their systems are prepared for stablecoin integration. This marks a transition from pilot programs to scalable implementations across treasury, risk management, and compliance functions.
Regulatory Environment Stabilizes
Perhaps most significantly, regulatory concerns have diminished dramatically. Only 18% of respondents now cite compliance or regulation as barriers, down from 80% two years ago. This shift reflects clearer policy frameworks, particularly in Europe with MiCA implementation, and improved AML tooling.
Regional Implementation Patterns
Latin America leads in practical deployment, with 71% using stablecoins for cross-border payments and 100% of surveyed firms either live or in planning stages. The region's focus on B2B import/export businesses highlights stablecoins' utility in trade finance.
Asia emphasizes market expansion, with 49% citing it as their primary driver. The region processes billions daily through global trade corridors, with 87% reporting technology readiness.
North America shows 39% adoption rates but 88% view upcoming regulations positively, suggesting accelerated implementation ahead. Companies like ALT 5 Sigma demonstrate the scale potential, growing from $39M to over $2B in transaction volume between 2020-2024.
Europe prioritizes security and systematic integration, with 58% using or planning stablecoin payments under the MiCA framework.
Infrastructure Requirements Crystallizing
The data reveals specific technical demands: 41% require high-speed processing, 34% demand robust compliance systems, and 36% identify enhanced security as crucial for scaling adoption.
Operational examples include Zeebu processing $5.7B in transactions across 139 telecom carriers, and traditional sectors like shipping and steel trading integrating stablecoin rails into existing workflows.
Fireblocks' latest survey of 295 financial executives reveals that stablecoin adoption has moved beyond experimentation into operational deployment, with 90% of respondents actively implementing or planning stablecoin payment systems.
Key Market Dynamics
The research shows a fundamental shift in priorities. Speed has emerged as the primary value proposition, cited by 48% of respondents, while cost savings ranked lowest at 30%. This suggests organizations view stablecoins as performance enhancers rather than cost-cutting tools.
Infrastructure readiness has reached a tipping point, with 86% of firms reporting their systems are prepared for stablecoin integration. This marks a transition from pilot programs to scalable implementations across treasury, risk management, and compliance functions.
Regulatory Environment Stabilizes
Perhaps most significantly, regulatory concerns have diminished dramatically. Only 18% of respondents now cite compliance or regulation as barriers, down from 80% two years ago. This shift reflects clearer policy frameworks, particularly in Europe with MiCA implementation, and improved AML tooling.
Regional Implementation Patterns
Latin America leads in practical deployment, with 71% using stablecoins for cross-border payments and 100% of surveyed firms either live or in planning stages. The region's focus on B2B import/export businesses highlights stablecoins' utility in trade finance.
Asia emphasizes market expansion, with 49% citing it as their primary driver. The region processes billions daily through global trade corridors, with 87% reporting technology readiness.
North America shows 39% adoption rates but 88% view upcoming regulations positively, suggesting accelerated implementation ahead. Companies like ALT 5 Sigma demonstrate the scale potential, growing from $39M to over $2B in transaction volume between 2020-2024.
Europe prioritizes security and systematic integration, with 58% using or planning stablecoin payments under the MiCA framework.
Infrastructure Requirements Crystallizing
The data reveals specific technical demands: 41% require high-speed processing, 34% demand robust compliance systems, and 36% identify enhanced security as crucial for scaling adoption.
Operational examples include Zeebu processing $5.7B in transactions across 139 telecom carriers, and traditional sectors like shipping and steel trading integrating stablecoin rails into existing workflows.
Fireblocks
Global Insights: Stablecoin Payments & Infrastructure Trends | Fireblocks
Stablecoins are transforming global payments. Discover 2025 key adoption drivers, regional trends, and why infrastructure will determine the winners.
ETH Zurich demonstrated a robot dog playing badminton using only onboard perception
In the video, a single RL policy coordinates 18 DOF simultaneously, achieving 10 consecutive rally shots with 12.06 m/s swing velocity and sub-400ms reaction time.
Video.
In the video, a single RL policy coordinates 18 DOF simultaneously, achieving 10 consecutive rally shots with 12.06 m/s swing velocity and sub-400ms reaction time.
Video.
Science Robotics
Learning coordinated badminton skills for legged manipulators
A legged mobile manipulator trained to play badminton with humans coordinates whole-body maneuvers and onboard perception.
❤2
Agent Zero is a personal agentic framework that dynamically grows and learns with you.
- It uses the OS as a tool.
- Has search and terminal execution too.
- It has persistent memory to memorize key information to solve future tasks more reliably.
- Multi-agent support.
- It uses the OS as a tool.
- Has search and terminal execution too.
- It has persistent memory to memorize key information to solve future tasks more reliably.
- Multi-agent support.
GitHub
GitHub - agent0ai/agent-zero: Agent Zero AI framework
Agent Zero AI framework. Contribute to agent0ai/agent-zero development by creating an account on GitHub.
🆒3❤1
Google presented Atlas (A powerful Titan): a new architecture with long-term in-context memory that learns how to memorize the context at test time.
Atlas even outperforms Titans, and is more effective than Transformers and modern linear RNNs in language modeling tasks.
It further improves the effective context length of Titans and scales to 10M context window with +80% accuracy on the BABILong benchmark.
Bonus: Building on Atlas ideas, researchers also discuss another family of models that are strict generalization of softmax attention.
Atlas even outperforms Titans, and is more effective than Transformers and modern linear RNNs in language modeling tasks.
It further improves the effective context length of Titans and scales to 10M context window with +80% accuracy on the BABILong benchmark.
Bonus: Building on Atlas ideas, researchers also discuss another family of models that are strict generalization of softmax attention.
🔥7