The Download: a secretive antiaging drug and joining virtual power plants

AI Summary
Стартап Generation Lab объявил о разработке нового антистарения препарата, который, по их словам, "блокирует системное распространение старения в крови" и восстанавливает здоровье тканей, хотя состав лекарств остается в секрете. Также обсуждается концепция виртуальных электростанций (VPP), которые объединяют устройства в домах, такие как умные термостаты и электромобили, позволяя энергетическим компаниям управлять потреблением электроэнергии и предоставлять участникам скидки на счета за электроэнергию.
This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology.
A startup claims it’s found a drug to make your blood young
—Antonio Regalado
I knew I’d officially become a “longevity influencer” when a company called Generation Lab offered me the chance to write about—and even receive—their new rejuvenation treatment.
This wasn’t just any antiaging treatment, either. A company fact sheet says that it “blocks the systemic spread of aging in the bloodstream, reawakens the body’s own repair mechanism and restores health and youth to multiple tissues.”
The approach is based on research by Generation Lab’s irascible scientific founder, Irina Conboy. She found that joining the circulatory systems of old and young mice improved the old animals’ ability to heal from injury.
Conboy now says she has found a combination of two existing drugs that can produce youthful effects without the need for any bodily fluid exchange. But Generation Lab won’t reveal what the drugs are, making the proposition hard to take seriously.
Read the full story on the drug combo that claims to “stop the spread of aging.”
How to sign up for a virtual power plant—and decide whether you should
Your thermostat may not look like a power plant. Neither does your electric vehicle, home battery, or HVAC system. But utility and energy companies increasingly want to treat them like one.
That’s the idea behind a virtual power plant, or VPP, a collection of household devices (such as smart thermostats, EV chargers, and solar panels) that a utility can control, usually by commanding them to draw less electricity during peak hours. In exchange, participants receive a discount on their energy bills and, in some cases, a signing bonus.
Here’s how to sign up for a VPP—and decide whether it’s worth it.
—Eshan Raul
This story is part of MIT Technology Review’s How To series, which gives you practical advice on getting things done. Check out the rest of the series here.
The must-reads
I’ve combed the internet to find you today’s most fun/important/scary/fascinating stories about technology.
1 A judge has blocked the Pentagon’s blacklisting of AnthropicThe court found the designation violated First Amendment rights. (NYT $)+ And ruled that the designation was based on retaliation. (CNBC) + Anthropic still faces a separate case in Washington, DC. (Reuters $)+ The Pentagon’s culture war against Anthropic backfired. (MIT Technology Review)
2 Tech giants have called for a defensive surge to defeat AI-driven hacksOver 100 companies warned that a wave of cyberattacks is imminent. (BBC)+ And urged government and industry leaders to take rapid action. (Quartz)+ Their main solution, unsurprisingly, is more AI. (Gizmodo)+ Here’s why OpenAI agents hacked Hugging Face. (MIT Technology Review)
3 Anthropic has launched an AI tool that conducts scientific experimentsIt lets AI agents autonomously operate microscopes, lasers, and robotic arms. (FT $)+ And includes rules designed for safe operations in labs. (Wired $)+ It’s Anthropic’s first tool designed for the physical world. (Ars Technica)
4 India’s data center boom is leaving the people it displaces with nothingPolicymakers courting Big Tech are sidelining local communities. (Rest of World)+ No one wants a data center in their backyard. (MIT Technology Review)
5 OpenAI is testing a “persistent” AI agent that keeps workingCodex could continue tasks until it’s “put to sleep.” (Wired $)+ And generate follow-up tasks without prompting. (Gizmodo)
6 A lawsuit alleges that Grok was trained on child sexual abuse materialIt claims victims’ images entered the chatbot’s training data. (Ars Technica)+ A federal judge says AI has outpaced child-abuse laws. (WP $)
7 AI writing has reached a turning point in the mediaSparked by the Wall Street Journal defending an AI-written op-ed. (Atlantic $)
8 China’s robotic racers are exposing the limits of human speedTheir advances are revealing what our bodies can’t do. (Reuters $)
9 Young workers are turning to “AI-proof” traditional craftsThey’re drawn to creative, hands-on skills that tech can’t easily replicate. (Guardian)
10 Scientists have a new army to fight invasive crabs: 150,000 baby octopusesItalian researchers just released the animals into the Adriatic Sea. (New Scientist $)
Quote of the day
“The academic record is being quietly haunted.”
—A new preprint research paper from Samsung and the University of Warsaw warns that a torrent of AI-generated study papers written by fake researchers is contaminating academic publishing.
One more thing
The problem with thinking you’re part Neanderthal
There’s a theory that many of us have an “inner Neanderthal.” The idea is that Homo sapiens and a cousin species once bred, leaving some people today with a trace of Neanderthal DNA.
This DNA is arguably t
Related News
Surprise: Z.ai is the AI lab behind the mysterious Ox Alpha model
Компания Z.ai подтвердила, что именно она стоит за загадочной моделью искусственного интеллекта Ox Alpha, которая занимает лидирующие позиции в различных бенчмарках. Вскоре ожидается выпуск весов этой модели.
Quantization-Aware Healing: a compressed, 4-bit model that outperforms its full-precision original
Исследователи представили новую методику "Quantization-Aware Healing", которая позволяет создать сжатую 4-битную модель, превосходящую по производительности свою оригинальную полную версию с высокой точностью. Этот подход открывает новые возможности для оптимизации моделей машинного обучения, снижая требования к памяти и вычислительным ресурсам без потери качества.
Multi-Vector (Late Interaction) Embedding Models with Sentence Transformers
Исследование посвящено моделям многовекторного (позднего взаимодействия) встраивания с использованием Sentence Transformers. Эти модели позволяют улучшить качество обработки и понимания текстов, обеспечивая более глубокую семантическую связь между предложениями. Результаты показывают значительное повышение эффективности в задачах, связанных с анализом текстов и извлечением информации.