islander wrote:
OIL was the "it" technology for over a century. Now we find it's irrevocable effect on the planet. The so-called National Security race for A.I. against China is absurd. They are already eating our lunch regarding EV's in advancement, production and cost. Hello, Prediction Markets — I'm betting by 2030, American Auto makers will beg for a bailout.

 miamizsun wrote:

Sort of true. But the current AI boom is like putting an oil well on every corner because we are planning to sell a gas guzzling Suburban to everyone in the next year. There are other solutions for transport, not everyone will buy a suburban, and if they did we probably still won't pump oil from all these new wells.  

Data and compute are certainly part of our new economy (Amazon makes more from compute than it does selling trinkets). The concern you are seeing is about the large resources being put into this (money, concrete, RAM, electrical infrastructure, water resource allocation), with an unknown return, and little or no benefit to most of the communities that are going to wind up dealing with them.

 islander wrote:
or make a res for a Ballroom Bunker Suite.

 miamizsun wrote:
 
Appreciate the discourse. Always ready to learn 
That 'rural' thang (sounds good to me) has run into a big Ruby Red snag in Texas, with MAGA ranchers, farmers and old landowners spittin nails. These good ol boys are talking about the blasphemous move of voting for a Democrat. Abbott has passed laws preventing local voices from having any rights or input regarding DC site construction. They are a pissed bunch.

 black321 wrote:

I was in the industry up until ~ 2017. We were a small provider, but we did some really innovative things. Our larger data center was very high power density and also super efficient. We worked closely with the local power provider and were the largest single contributor to their conservation programs for 3 years in a row.  I spent a LOT of money being green. I did it partly because I'm a nerd and I don't want to polute the planet. But it was an easy economic sell - typical ROI on our projects was 3-4 years, and it was usually 20 year lifecycle equipment. 

The way the current builders waste money is what worries me now. The green impacts are legitimate too, but the simple shortsightedness is odd. For a bunch of smart guys that are infatuated with money, they all seem to be placing big bets, and not all of them will win.  There are lots of ways to do it differently, make it more sustainable and have a long term view. But these guys just go out and swing for the fences. I think most of them have already insulated themselves personally from any failures. So why not go big, and try to be the next trillionaire? if it fails, you will just shuffle the failure off to the shareholders, or community,  or go ask for a bailout.

 miamizsun wrote:

I dont disagree....but "perhaps" the reason for the growing concern is the level/pace of recent investment, chasing Ai. 
eg, Google, Meta, Microsoft and Amzn alone are investing ~$1T just this year.

An Amazon center: https://images.wsj.net/im-2966...

TL;DR: Today’s top "AI for Good" developments feature the launch of a major open-science biological design initiative led by a Nobel laureate, digital twin modeling for virtual clinical trials, and fresh frameworks for evaluating AI outcomes in public service.

Key AI for Good Breakthroughs

• Open-Science "AI BioDesign" Project Launches to Rewrite Biological Rules: The Allen Institute, University of Washington, and Fred Hutch Cancer Center formally launched AI BioDesign. Led by Nobel laureate David Baker, the collaborative open-science platform pairs generative AI models with high-throughput synthetic biology. The initiative focuses on generating open models, assays, and datasets to design custom biological tools—ranging from targeted cancer therapeutics and enzymes that degrade ocean microplastics to bio-based, low-power computing components. Read the full announcement from >"="" href="https://newsroom.uw.edu/news-releases/ai-biodesign-project-aims-to-adapt-natures-design-rules" data-hveid="0" decode-data-ved="1" data-ved="0CAAQ_4QMahgKEwiMtOWRm9WWAxUAAAAAHQAAAAAQ8gI">UW Medicine Newsroom.
• Multimodal AI "Digital Twins" Enable Virtual Clinical Trials: Research presented at the Fred Hutch AI Club detailed how multimodal generative AI models (such as PubMedBERT and BiomedParse) are constructing virtual patient "digital twins" from longitudinal electronic health records, imaging, and multiomics data. By simulating disease progression and population-scale drug interactions, the framework enables virtual clinical trials that cut traditional multi-million-dollar R&D timelines while personalizing precision health pathways. Explore the session background on >"="" href="https://www.fredhutch.org/en/events/ai-club-sept-2026.html" data-hveid="0" decode-data-ved="1" data-ved="0CAAQ_4QMahgKEwiMtOWRm9WWAxUAAAAAHQAAAAAQ8wI">Fred Hutch Events.
• Global Governance Shift Focuses AI Evaluation on Tangible Public Outcomes: Insights from global AI governance sessions in Geneva highlight a crucial shift in deploying AI across public health, agriculture, and social safety nets. UN bodies and research organizations are codifying frameworks that move past static vendor benchmarks, establishing independent, continuous evaluations to ensure AI tools deliver measurable real-world outcomes—such as symptom resolution in community health—without performance degradation over time. Read the governance breakdown on >"="" href="https://www.idinsight.org/article/governing-ai-takeaways-from-the-2026-ai-for-good-global-summit/" data-hveid="0" decode-data-ved="1" data-ved="0CAAQ_4QMahgKEwiMtOWRm9WWAxUAAAAAHQAAAAAQ9AI">IDinsight.

 Isabeau wrote:

There were data centers long before Trump and they will be here long after he's gone. Trump in all of foolishness is irrelevant. The last time I looked, the USA has over 4,000 data centers. Without them life would cease as we know it. Literally everything we use is touched by electricity, data and compute. Earth needs more, not less. There seems to be a flare up of data center concern and/or paranoia lately. Why? If we moved them to rural areas and had behind the meter energy generation with solar, nat gas or better yet nuclear would that be acceptable?

Regards

Miamizun wrote: Half of the dedicated capital is directed specifically toward "Green AI"—developing model architectures and training pipelines focused on reducing algorithmic energy usage and carbon footprints, alongside hybrid models designed for environmental modeling.
___________________________________________________________

Unfortunately, that will not be the immediate future and certainly not here in the U.S. if Trump mentality reigns. The AI DCs that are already functioning, those being built and those about to go online are NOT all loop cooling systems or utility regulated. The constant humming, the light pollution, emissions from back up generators are still en route to average American's backyards.
Texas is experiencing a massive data center boom with over 300 operating facilities, 200 planned projects and nearly 100 under construction. 

The decimation of US geography and communities is the price we pay for this alleged 'good' from AI? 

gleaning a biotech conference material and the thing that everyone wants is more AI/compute
seems cancer is at the top of most hit lists...

TL;DR: Today’s top "AI for Good" developments feature a 1,000x cost reduction in virtual drug discovery published in Nature Biotechnology, a collaborative global cancer initiative from Harvard, MIT, and Johns Hopkins, and a European public funding framework dedicated to sustainable "Green AI."

Key AI for Good Breakthroughs

• Open-Source Platform Screens 69 Billion Molecules for Cancer: Researchers from St. Jude Children’s Research Hospital, Harvard Medical School, and the Dana-Farber Cancer Institute launched AdaptiveFlow, an open-source AI platform published in Nature Biotechnology. The framework achieves a 1,000-fold reduction in computational costs for ultra-large drug screenings, allowing researchers to screen a world-record 69 billion virtual molecules to identify potent inhibitors for challenging cancer targets in hours rather than months. Read the full study announcement on >"="" href="https://www.stjude.org/media-resources/news-releases/2026-medicine-science-news/ai-informed-adaptiveflow-redefines-large-scale-cloud-computing-for-drug-discovery.html" data-hveid="3">St. Jude News.
• AI-Enabled "Cancer Center in the Cloud" Launches: Global Health Catalysts unveiled the Comprehensive Cancer/CARE Center in the Cloud (C4), an AI-powered ecosystem co-developed by researchers at Harvard, MIT, Johns Hopkins, and the University of Pennsylvania. Designed to bridge global healthcare disparities, the platform connects U.S. clinical expertise, AI diagnostic agents, and real-time research with health workers in underserved international communities. Read the project rollout on >"="" href="https://www.einpresswire.com/article/937722705/global-health-catalyst-summit-2026-to-launch-ai-enabled-cancer-center-in-the-cloud-at-harvard-mit" data-hveid="4">EIN Presswire.
• European Union Ecosystem Expands "Green AI" Funding: European research bodies announced new major funding initiatives under the "AI Ecosystems 2026" call. Half of the dedicated capital is directed specifically toward "Green AI"—developing model architectures and training pipelines focused on reducing algorithmic energy usage and carbon footprints, alongside hybrid models designed for environmental modeling. Read the program details on >"="" href="https://www2.fundsforngos.org/community-development-2/ai-ecosystems-2026-hybrid-ai-and-green-ai-funding-call-austria/" data-hveid="5">FundsforNGOs.
• ARPA-H $98.5M AI Rare Disease Program Enters Operational Phase: Following its recent award allocations, the Advanced Research Projects Agency for Health (ARPA-H) is onboarding participating medical centers to aggregate unified genomic and EHR datasets. The privacy-preserving AI network aims to systematically shorten the multi-year diagnostic path for millions of rare disease patients worldwide. Read the official update on >"="" href="https://arpa-h.gov/news-and-events/arpa-h-announces-awards-advance-ai-rare-disease-diagnosis-and-treatment" data-hveid="6">ARPA-H.

According to the article, artificial intelligence and machine learning are being used to accelerate discovery by analyzing complex biological datasets, modeling molecular interactions, and helping scientists move from understanding biology to designing therapeutics more efficiently.

What comes after understanding cancer biology? New ways to act on it

Key Applications of AI in the Pipeline

• Data Analysis: AI algorithms parse through increasingly complex biological datasets that would be exceedingly difficult to analyze manually.
• Molecular Modeling: Machine learning helps model complex molecular interactions, allowing researchers to generate stronger hypotheses regarding disease-driving targets.
• Therapeutic Design: These computational tools shorten the timeline between biological discovery and the development of actionable modalities like targeted protein degradation.

 Isabeau wrote:

Should be abolished and replaced with universal medical care.

 Red_Dragon wrote:
or in handling health insurance claims.

 miamizsun wrote:

Scientific and medical research is where is belongs - not in creativity.

Health and Science

Cancer vaccine from Moderna, Merck shows promise in late-stage trial; both stocks soar

Published Wed, Aug 19 20266:46 AM EDT Updated 2 Hours Ago


Key Points

• An experimental personalized cancer vaccine from Merck and Moderna showed positive initial results in its first-ever late-stage trial.
• The combination regimen met the main study goal of significantly extending the time patients lived without their melanoma returning, compared with Keytruda alone, and also reduced the risk of the cancer spreading to distant parts of the body.
• The initial data suggests that the combo could serve as a new treatment option for the deadliest form of skin cancer, and validate a new personalized treatment approach.

J, while public discourse often fixates on the disruptive risks of artificial intelligence, its empirical application is already driving measurable improvements in fields like medicine and resource management. Machine learning algorithms excel at identifying complex patterns within massive datasets, accelerating practical breakthroughs such as the design of personalized cancer therapeutics. By treating AI as a powerful analytical tool rather than a speculative threat, researchers are actively deploying it to solve previously intractable human challenges.
- Rex