Here is the daily technology #threadcast for 5/16/25. The goal is to make this a technology "reddit".
Drop all question, comments, and articles relating to #technology and the future. The goal is make it a technology center.
Here is the daily technology #threadcast for 5/16/25. The goal is to make this a technology "reddit".
Drop all question, comments, and articles relating to #technology and the future. The goal is make it a technology center.
CRYPTOPUNKS JOIN THE NODE FOUNDATION
The Infinite Node Foundation has acquired CryptoPunks IP from Yuga Labs, placing the pixel-art collection in a mission-driven nonprofit endowment to safeguard its long-term cultural legacy. NODE will pursue preservation of blockchain integrity, foster a community ecosystem, and expand contexts for study and exhibition under a networked stewardship model. It plans to create a permanent Palo Alto hub with a full Ethereum node and a debut exhibition of all 10,000 Punks to integrate this pioneering digital art into broader art-historical discourse.
Are We Ready for 100 Million Onchain Users
The onchain user experience remains fragmented and friction-filled. 48% have multiple wallets, transactions require an average of five approvals and three gas fees, and despite 54% having used crypto for payments, only 12% call it their favorite use case. While 69% of users report feeling safe onchain, 21% have been phished in the past year. Chains continue to specialize, with Ethereum for settlements, Solana for speed, Base for experimentation, and BNB for exchange-centric services. To onboard the next 100 million users, chain complexity must be abstracted, and security measures must be improved. There should be user interfaces for novice and professional users.
Best Bitcoin Yield Opportunities
Bitcoin yield opportunities range from 4% to over 20% APR across diverse DeFi strategies with varying risk levels. Liquidity-mining opportunities, such as Solv Protocol with Infrared Finance or Velodrome offer stable rewards for BTC liquidity. Funding rate arbitrage and looping primitives like Euler's farming on cbBTC/LBTC or Pendle-powered Morpho loops unlock higher yields via leveraged positioning. Cash/spend tools and simple leverage platforms, including EtherFi's Cash Card and Fluid's smart collateral loops, round out a spectrum of capital-efficient, composable yield options.
Introducing Multiverse Finance
Multiverse Finance has proposed partitioning the financial system into parallel “verses” tied to event outcomes, allowing users to collateralize and borrow conditional tokens (e.g., notFiredUSD vs. notFiredETH) without liquidation risk by pushing and pulling ownership across these verses. The framework uses a “multiverse map” primitive to track balances per verse and supports splitting and combining operations that mirror prediction market minting and redemption. Upon resolution via an oracle, all but the realized verse evaporate, enabling seamless withdrawal or settlement. This design promises fully composable, multiverse‑aware DeFi applications and generates predictive information about real‑world events as a valuable byproduct.
Ethereum stablecoin volume hits record $908 billion as institutions, tech giants, and Trump jump in
Ethereum-based stablecoins saw a record $908 billion in transaction volume last month driven by rising institutional adoption, tech giant integration, and increased political engagement with crypto. Analysts attribute the surge to major fintech and payment players expanding support for stablecoins on Ethereum, alongside high-profile political endorsements that have reframed digital assets as a campaign issue.
Microsoft’s CEO on How AI Will Remake Every Company, Including His
AI is rapidly commoditizing, and this is a good thing for Microsoft. CEO Satya Nadella believes that as AI becomes more accessible and its usage becomes more efficient, consumption will increase. The company has created its own AI architectures to prepare for this future. Microsoft's investment into AI projects will likely generate trillions of dollars in economic activity.
MoonPay and Mastercard Partnership
MoonPay and Mastercard have partnered to let every crypto wallet issue virtual Mastercards that draw directly from stablecoin balances, enabling payments at over 150 million merchants worldwide. This integration significantly expands stablecoin utility in real‑world transactions and aligns with Mastercard's end‑to‑end stablecoin capabilities. It follows similar crypto‑card rollouts with OKX and Kraken, highlighting the payments giant's deepening engagement in digital asset flows.
From kneecaps to spaceships: Explainable AI can help create superalloys faster than ever
The future of material design is here. A unique AI system can help scientists create perfect alloys faster than ever.
Can you guess the recipe for a cake just by looking at a bakery shelf filled with hundreds of finished desserts? That’s the kind of challenge materials scientists have faced for decades, trying to guess the elements that combine well and result in useful materials.
With over 100 elements on the periodic table, there are an endless number of ways in which these elements could potentially combine. Scientists can’t test each such combination, and therefore, it is very challenging to create materials and alloys with desirable properties.
A new study reveals an AI-based technique that can solve this problem. The AI can rapidly check countless possibilities, predict the most promising element combinations, and even suggest the ideal proportions of each element for creating alloys with superior properties.
Using this approach, the study authors produced multiple principal element alloys (MPEA) with exceptional mechanical properties. “This work demonstrates how data-driven frameworks and explainable AI can unlock new possibilities in materials design,” Sanket Deshmukh, one of the study’s authors and a professor at Virginia Tech, said.
!summarize #genx #robotics #ai
CarPlay Ultra's first trial reveals a deeply integrated, Apple-like experience
Apple's CarPlay Ultra is now available. Shipping initially in the Aston Martin, the next-generation interface expands the iPhone-based interface across multiple displays on the dashboard. It is able to display elements such as speed and fuel level. The new CarPlay Ultra requires an iPhone 12 or later running on iOS 18.4 or newer. A video that examines the new features of CarPlay Ultra is available in the article.
Coinbase says hackers bribed staff to steal customer data and are demanding $20 million ransom
Coinbase customer data was recently stolen by cybercriminals using social engineering techniques. The incident may cost the company up to $400 million to fix. The data stolen includes names, addresses, phone numbers, emails, masked bank account numbers and identifiers, the last four digits of Social Security numbers, government ID images, and account balances. The attackers demanded a $20 million ransom, but Coinbase said it would not pay. It instead offered a $20 million reward fund for information leading to the arrest and conviction of the criminals responsible for the attack.
A Gene-Editing Breakthrough Saves Infant With Rare Disease
An infant in Philadelphia has been successfully treated with a customized CRISPR gene-editing therapy for a rare and deadly disease. The infant lacked an enzyme in his liver that normally converts ammonia to urea. This caused ammonia to build up and damage organs in the body. Before the treatment, the infant was in the eighth or ninth percentile for weight - after, he is now in the 35th to 40th percentile. The infant still needs medication, but now needs less. His treatment could potentially be used as a template for creating individualized treatments quickly for other patients with rare diseases.
For the first time in the US, a rotating detonation rocket engine takes flight
Venus Aerospace has completed a short flight test of its rotating detonation rocket engine. The concept has been discussed academically for decades and has previously been tested in a handful of countries, but this is believed to be the first US-based flight test of the idea. The technology makes traversing the globe in under two hours possible. Venus still has a long way to go before it can provide highly efficient hypersonic commercial flight, but the flight test proves that such a future is at least possible.
Working on Complex Systems
Understanding the difference between complicated and complex problems is important because each requires a fundamentally different approach. Complicated problems are intricate but predictable, following structured, repeatable solutions, while complex problems are unique, requiring adaptive and often novel solutions. Recognizing whether a system is complicated or complex can help lead to solutions with more effective results. In many environments, systems are neither purely complicated nor purely complex - the key is learning to recognize when adaptability is needed and when structural solutions are enough.
Meta Is Delaying the Rollout of Its Flagship AI Model
Meta's engineers are struggling to significantly improve the capabilities of the company's 'Behemoth' large language model. The model was intentionally slated for an April release, but this has now been delayed to fall or later. Staff are concerned that the model's performance won't match public statements about its capabilities. The company is contemplating significant management changes to its AI product group.
Stack overflow is almost dead
The number of monthly questions on Stack Overflow this month is as low as when the site launched in 2009.
!summarize
#ai
Things are not going according to plan at Meta AI: According to the Wall Street Journal, the social media giant is postponing the release of 'Behemoth', its flagship AI model, as engineers struggle to deliver any notable improvements over previous versions. The launch has been pushed to later this year, as concerns ramp up around the model’s performance gains. The delay has reportedly created internal friction, with senior executives "frustrated" at the team's progress and contemplating major changes.
https://www.wsj.com/tech/ai/meta-is-delaying-the-rollout-of-its-flagship-ai-model-f4b105f7?utm_source=superhuman&utm_medium=newsletter&utm_campaign=meta-s-struggles-and-openai-teases-next-release&_bhlid=3313f7616d33442ec3899ebb081bac27c035c9b9
Viral startup Manus AI unveils new image gen capabilities: Unlike single-purpose AI tools that just whip out images, Manus’ new image generation tool doesn’t just adhere to text prompts — it understands your intent, plans workflows, browses the web for information, and pulls together multiple specialized tools to deliver ready-to-use visual content for marketing, product design, and other areas. You can see it in action here.
~~~ embed:1923048495310922028?utm_source=superhuman&utm_medium=newsletter&utm_campaign=meta-s-struggles-and-openai-teases-next-release&_bhlid=07a5cca3893fb75e73d2dee086a3dd12c8777864 twitter metadata:TWFudXNBSV9IUXx8aHR0cHM6Ly90d2l0dGVyLmNvbS9NYW51c0FJX0hRL3N0YXR1cy8xOTIzMDQ4NDk1MzEwOTIyMDI4fA== ~~~
OpenAI teases its next big release: The ChatGPT-maker has social media on edge after posting a cryptic teaser hinting at a new AI tool for developers. CEO Sam Altman kicked off the teaser campaign yesterday evening, promising "another low-key research preview" that will have a better name than ChatGPT "incase it takes off.
~~~ embed:1923212267677499469?utm_source=superhuman&utm_medium=newsletter&utm_campaign=meta-s-struggles-and-openai-teases-next-release&_bhlid=a568126a5fca4ba3e58adf7f1f13cb6ca4630f7a twitter metadata:T3BlbkFJfHxodHRwczovL3R3aXR0ZXIuY29tL09wZW5BSS9zdGF0dXMvMTkyMzIxMjI2NzY3NzQ5OTQ2OXw= ~~~
Done Deal: The UAE has inked a deal with the US to build the world's largest AI campus outside America — a huge facility that could potentially support 2.5M of Nvidia's top-line chips. It’s a massive push for the country’s AI ambitions.
https://www.reuters.com/world/china/uae-set-deepen-ai-links-with-united-states-after-past-curbs-over-china-2025-05-15/?utm_source=superhuman&utm_medium=newsletter&utm_campaign=meta-s-struggles-and-openai-teases-next-release&_bhlid=4350843b909469e0127dcc0dc057b74bee39c029
Code Cruiser: Vibe-coding startup Windsurf has released its first in-house AI model SWE-1, optimized for "the entire software engineering process." It reportedly performs on par with Claude 3.5 Sonnet and GPT-4.1 on internal benchmarks, but falls short of Claude 3.7 Sonnet.
https://techcrunch.com/2025/05/15/vibe-coding-startup-windsurf-launches-in-house-ai-models/?utm_source=superhuman&utm_medium=newsletter&utm_campaign=meta-s-struggles-and-openai-teases-next-release&_bhlid=d1fee10080bca8b26c5ff98bd2017923cb58e68e
Dr. GPT: Open-source startup Intelligent Internet Inc. just announced a medical AI model that claims to match GPT 4.5 while running locally on any modern laptop — at no cost.
~~~ embed:1922998848286286209?utm_source=superhuman&utm_medium=newsletter&utm_campaign=meta-s-struggles-and-openai-teases-next-release&_bhlid=549a7a91391f0c46b20d34dcd77fd49ded6da189 twitter metadata:aWlfcG9zdHN8fGh0dHBzOi8vdHdpdHRlci5jb20vaWlfcG9zdHMvc3RhdHVzLzE5MjI5OTg4NDgyODYyODYyMDl8 ~~~
Confession of a Bot: Instead of asking ChatGPT to be the therapist, one user flipped roles and asked ChatGPT to share its problems. The answers are fascinating.
Shots fired: Sam Altman took a swipe at Grok, and Lovable CEO Anton Osika called out Replit CEO Amjad Masad. Things are getting heated.
Agent Academy: Builder Santiago shared his roadmap to learning how to build AI agents.
UK scientists solve puzzle of loud next-gen aircraft as EU eyes quieter skies
The research will aid in design of quieter aircraft to be used for large scale air transport as well as eVTOLs.
A collaboration of researchers at the universities of Bristol and Salford in the UK has identified the cause that results in an irritating noise produced by next-generation engines that will be used to build hybrid and electric aircraft in the future. The discovery paves the way for airplanes that not only sound quieter but are also quiet when engine noise is measured.
The noise produced by jet engines is well known, and its impact on our health is well understood. Efforts are ongoing to reduce noise output from these engines. However, as we transition to a new era of aerial transportation powered by cleaner energy sources, the associated engine technology has also been developed to ensure that this revolution is quieter.
At the outset, electric engines come across as much quieter counterparts of the jet engine. While this is partly true, the noise produced by an electric engine is produced at different frequencies from their jet engine counterparts, which makes it sound quieter.
Tiny robots team up like humans with 84% mission success, learn to ambush, encircle
HUMAC capitalizes on the uniquely human ability to grasp and anticipate others’ intentions.
Researchers from Duke University and Columbia University have successfully harnessed a uniquely human cognitive ability – the “Theory of Mind” – to rapidly train teams of robots to perform complex collaborative tasks.
“Humans start to develop the skill of Theory of Mind around age four,” explained Boyuan Chen, the Dickinson Family Assistant Professor at Duke University.
“It allows us to interpret and predict others’ intentions, allowing collaboration to emerge. This is an essential capability that our current robots are missing to allow them to work as a team with other robots and humans.”
The new framework, dubbed HUMAC (Human-guided Multi-Agent Collaboration), allows a single human coach to impart sophisticated teamwork skills to robots by intuitively guiding them through key strategic moments.
“We designed HUMAC to help robots learn from how humans think and coordinate in an efficient way,” added Chen.
!summarize #ericschmidt #agi #ai #biotech
Paragraphs
The study authors focused on MPEA, special alloys composed of more than three elements (in roughly equal proportions) that exhibit remarkable mechanical properties. These materials are much stronger, tougher, and more resistant to heat and wear than conventional alloys.
They are already used in aircraft parts, surgical implants, and clean energy systems. However, the traditional method of creating MPEA involved testing numerous combinations of elements. This process is costly and time-consuming.
To overcome this challenge, Deshmukh and this team developed a smart, data-driven system that combines the power of machine learning and algorithms. They started by gathering a large dataset of existing MPEA.
Basically, information like which metals were used, how those metals were arranged at the atomic level, and how each alloy performed in terms of strength, flexibility, and other mechanical properties was fed to the system.
Two machine learning models were trained using the information. The first was a Stacked Ensemble Machine Learning (SEML) model, which used the composition (the elements and their amounts) to predict how strong or flexible an alloy would be.
The second model was a Convolutional Neural Network (CNN), a type of AI usually used for analyzing images, but here, it was used to examine how atoms of different elements were arranged next to each other inside the alloy, as this structure has a big impact on how the metal behaves.
Once these models could predict properties accurately, the team used algorithms to improve the results. These algorithms generated many possible metal combinations, kept the best ones, and mixed them to create even better versions. Over time, the system was able to zoom in on the most promising alloy designs.
“Our design workflow, combining advanced machine learning and evolutionary algorithms, provides interpretable insights into materials’ structure-property relationships, offering a robust approach for the discovery of diverse advanced materials,” Fangxi Wang, first author and a postdoc researcher at Virginia Tech, said.
The researchers are confident that their work will lead to the design of embedded engines that not only sound quieter but are also quiet when measured.
“By linking turbulent flow ingestion patterns to how people perceive noise, we are giving engineers the tools to design future aircraft that truly sound as quiet as they look,” added Ahmed in the press release.
The efforts are crucial in light of the European Union’s mandate to reduce aircraft noise by 65 percent by 2050. The researchers believe that their work will aid in design of quieter aircraft to be used for large scale air transport as well as for electric vertical take-off and landing (eVTOL) aircraft in the urban air mobility (UAM) sector.
When the engines operate at high thrust, during take-off, the fan suction is much stronger and disrupts the airframe boundary layer flow. This produces an extremely unsteady turbulent flow alongside fan-induced flow distortion. The interaction between fan blades and the distorted air flow results in fan haystacking, where the large span of the rotating blades further slices the unsteady flow.
“These two hidden sound signatures make future embedded aircraft engines feel perceptually irritating, not just loud,” explained Feroz Ahmed, who was involved with the work while at the University of Bristol in a press release.
Novel approach to robot training
This approach is different from traditional methods that often rely on extensive trial-and-error learning or costly imitation of large groups of human experts.
HUMAC leverages the human capacity to understand and predict the intentions of others. It enables robots to learn to anticipate their teammates’ actions, adapt strategies dynamically, and solve challenges demanding coordinated, collective intelligence akin to human teams.
The HUMAC framework involves brief instances where a human operator takes control of individual robots within a team during training, offering guidance at crucial strategic junctures – much like a coach providing targeted advice during a game.
These short demonstrations teach the robots intricate collaborative tactics, such as setting up ambushes and encircling targets.
Babbling Bot: A glitch in xAI’s Grok caused it to sprout off controversial “conspiracy theories” in response to unrelated prompts. Now, xAI has offered an explanation, blaming the glitch on an “unauthorized modification” and promising new safety checks.
Pay Day: Hedra, the AI startup that went viral recently for clips of talking baby podcasts, has picked up $32M from a16z to make its video avatars more expressive.
https://techcrunch.com/2025/05/15/hedra-the-app-used-to-make-talking-baby-podcasts-raises-32m-from-a16z
Prof GPT: A school in Texas is reportedly using AI instead of human teachers to educate children. The report claims students are scoring in the top 2% nationally.