UN Security Council AI briefing puts OpenAI, Anthropic, DeepSeek, and Moonshot in one room for the first time. AI safety enters diplomacy.
Get the latest international news and world events from around the world.
Xiaomi made an iPhone Duo. It’s Unhinged
Can the Xiaomi 18 Fold take on the iPhone Duo? 😅
Go to https://surfshark.com/boss or use code BOSS at checkout to get 4 extra months of Surfshark!
For my tech hot takes: / mrwhosetheboss.
For my Personal Posts: / mrwhosetheboss.
For quick tech videos: / mrwhosetheboss.
Does anyone still use this anymore?: / mrwhosetheboss.
My 5 favourite pieces of Tech right now (Not sponsored but these are affiliate links which help out the channel!):
1. Balmuda Toaster: https://geni.us/4B5xooB
2. Ember Mug 2: https://geni.us/zb4Ix.
3. Ninja Creami Deluxe: https://geni.us/Mseu.
4. Anker PowerBank: https://geni.us/p3PVF
5. Momax Travel Adaptor: https://geni.us/TJg5kGD
Music is from Epidemic sound:
http://share.epidemicsound.com/pHDFT
AI & the Threats & Dangers of Data Poisoning
Chuck Brooks is the president of Brooks Consulting International and one of Executive Mosaic’s GovCon Experts.
In my book Inside Cyber, I explored how artificial intelligence has evolved into the most potent weapon in an attacker’s arsenal as well as our most effective defensive tool. This dual nature is at the core of a significant change in security and privacy. The rate of change has only quickened since the book was released. We have advanced further into what I refer to as the Acceleration Era, in which AI systems, including large language models, or LLMs, improve quickly, incorporate into crucial processes and have a growing impact on business, governmental and societal decisions.
Data poisoning, or the intentional or unintentional contamination of the training data that forms these models, is one of the most pernicious new threats in this setting. Recent studies have highlighted LLMs’ continued vulnerability. A surprisingly small number of carefully constructed malicious documents, roughly a few hundred—can implant backdoors or change behavior in models with hundreds of millions to billions of parameters, according to studies, including collaborative work involving Anthropic, the UK AI Security Institute and the Alan Turing Institute. The amount of poisoned material does not have to increase in proportion to the size of the model or the amount of training data. The malicious samples’ absolute presence is what counts.
Alibaba Brags That Its Next-Gen Zhenwu V900 AI Chip Offers More On-Package Memory Than NVIDIA H200, Outlines Plans To Deploy 20 GW Of Compute, And Teases 4–10 trillion parameters For Upcoming Qwen Models
Alibaba is now claiming that the upcoming Zhenwu V900 chip will have 216GB of on-package memory, with an official launch slated for Q1 2027. We can only theorize that the V900 will leverage CXMT’s HBM3E solution, especially given their overlapping volume production timelines. The accelerator will sport chip-to-chip interconnect speeds of around 1.2 TB/s via Alibaba’s ICN Switch fabric, and offer around 3x the performance of Zhenwu M890, replete with native FP8/FP4 support. This means that each accelerator will offer a peak computing power of around 1.8 PFLOPS at FP16, given the ~0.6 PFLOPS that M890 had claimed to offer.
Critically, the ICN Switch can allow around 1,000 Zhenwu V900 chips to function as a single accelerator. However, Alibaba is now claiming that each V900 cluster can scale to 500,000 chips, entailing a whopping 108 petabyte of memory across the entire cluster! It remains to be seen if CXMT can fulfill the entirety of this oncoming demand.
Also, Alibaba is now offering its own bespoke rack-scale solution, replete with Yitian CPUs, Zhenwu V900 GPUs, ICN interconnect, Pangu NICs, and Zhenyue storage controllers.
Largest dataset of its kind could clarify how massive stars shaped early galaxies
The more astronomers learn about the universe’s earliest galaxies, the stranger they seem. Many of their surprising properties may be explained by differences between their massive stars and those in galaxies like our own Milky Way. A new University of Utah-led survey with the Hubble Space Telescope is shedding light on the stellar astrophysics operating in early galaxies.
The survey, called the Treasury of Extremely Metal-Poor O Stars (TEMPOS), uses ultraviolet (UV) observations from Hubble’s Cosmic Origins Spectrograph (COS) to study massive stars in nearby galaxies that are the best available analogs of stars in the early universe.
The unprecedentedly large dataset from TEMPOS could help astronomers build better models of massive stars to understand how they shaped galaxies when the universe was young. Such models are essential to interpret observations of early galaxies now coming from the James Webb Space Telescope, which launched in 2021.
Cancer-related DNA changes may appear in blood years before diagnosis
Researchers found that genome-wide plasma cfDNA methylation profiles collected before diagnosis contained early molecular signatures associated with future prostate and breast cancer, with prostate silencer methylation providing the stronger risk stratification. Breast enhancer signals were weaker and varied by subtype and disease stage, indicating that cfDNA methylation may have greater value for risk assessment and surveillance than as a standalone diagnostic test.
Series of reactions reveals how complex carbon chemistry can begin in frigid space
Before there was Earth, there was chemistry. In the unimaginable cold, dark clouds where stars and planets are born, carbon molecules assemble into more complex forms—starting a chemical journey that could eventually deliver some of life’s basic ingredients to young planets. There’s only one problem: These molecules need heat to form. At least that’s what scientists have long believed. But new research from FIU chemist Alexander Mebel reveals such chemistry can actually happen at temperatures nearing absolute zero, which is −460°F (−273°C).
On Earth, these complex carbon molecules are usually associated with extremely hot temperatures, such as those associated with combustion. It has long been presumed that heat is necessary for these molecules to assemble in space. But when astronomers detected one of these molecules last year, it was found in the cold, dark region known as Taurus Molecular Cloud-1. This region, where solar systems begin, is anything but hot. Temperatures there hover around −443°F (−264°C).
Mebel’s research, published in The Journal of Physical Chemistry A, offers a possible answer as to how these molecules formed there and could improve understanding of how worlds like Earth can form.
What happens when a particle breaks apart? | Cenke Xu (UCSB)
The purpose of these Blackboard Talk lunches is for the science of one program to be explained to the other KITP program participants. These talks lead to cross program communications that work towards our goal of maintaining the unity of theoretical physics.
Connect with KITP everywhere:
https://twitter.com/@kitp_ucsb.
https://kitp-ucsb.bsky.social.
▶ / @kitp_ucsb.
/ kavli-institute-for-theoretical-physics.
The position of the KITP is that ownership and copyright of all online material — slides, text, audio, video, and podcasts — belongs to the author. KITP is providing dissemination for these materials but does not claim ownership. Any person citing these materials for scholarly purposes should provide an appropriate scholarly reference.