NVIDIA Engineer Raises 5 Million for AI Infrastructure

Key Takeaways

• <ul><li>A 73-megawatt AI data center called RACC is being built at Texas A&M's RELLIS Campus by ThisWay Global</li><li>Former NVIDIA engineer Alexander Tsado founded Udu Technologies to offer AI computing under $1/hour in six African countries</li><li>29% of employed job seekers quit AI interviews versus 14% of unemployed candidates, per a 6,731-person survey</li><li>WeRateDogs founder Matt Nelson bans all AI content and the charity has raised nearly $5 million for needy dogs</li><li>NVIDIA's FOX Blueprint at Siemens' Erlangen factory enables up to 20% cost savings through AI agent orchestration</li><li>MIT research says human owners must retain control over AI agents, including configuration, monitoring, and incident response</li><li>Mural Arts Philadelphia faces backlash after using AI-generated imagery in public art banners by artist Kelley Prevard</li><li>A new machine learning framework from Deepak Kumar and Nick P. Bassill improves urban heat island mapping for climate planning</li><li>PMLR provides open-access peer-reviewed machine learning research from conferences like ICML and AISTATS via proceedings.mlr.press</li><li>AI-assisted literature reviews require human oversight and transparent reporting per the RAISE framework and PRISMA 2020 guidelines</li></ul>

AI data centers expand in Texas with new RACC facility

A 73-megawatt data center is being built on 25 acres at Texas A&M's RELLIS Campus in Bryan, Texas. The facility is called the Research and AI Compute Center, or RACC. It is being developed by Austin-based ThisWay Global. The center will focus on sovereign artificial intelligence infrastructure.

Nvidia veteran brings AI computing to Africa through Udu Technologies

Alexander Tsado left NVIDIA, where he helped build early cloud computing infrastructure, to found Udu Technologies. The Nigerian entrepreneur aims to bring affordable AI computing power to six African countries. Udu offers cloud GPU access for less than $1 an hour and computing credits starting at $50 and $100. The company also trains local teams to manage AI infrastructure and works with governments and institutions across the continent.

Employed job seekers quit AI interviews at high rate

A survey of 6,731 job seekers found that 29% of employed candidates quit AI interviews, compared to 14% of unemployed candidates. The survey covered the UK, US, Canada, the European Union and Latin America. About 58% of people who completed an AI interview said they would refuse one if they had other job offers. The data suggests AI interviews work best on people who feel they have no other choice.

PMLR guide explains open-access machine learning research

PMLR, or Proceedings of Machine Learning Research, is an open-access series that publishes peer-reviewed research from major AI conferences. It covers conferences including ICML, AISTATS, COLT, and CoRL. The series is hosted at proceedings.mlr.press and managed by an editorial board of computer science researchers. It was formerly known as JMLR Workshop and Conference Proceedings and was rebranded in 2017.

Machine learning framework improves urban heat island mapping

Researchers Deepak Kumar and Nick P. Bassill developed a framework that uses machine learning to detect urban heat islands and support climate resilience planning. The study was published in Discover Life. The framework links AI models to governance and biodiversity preservation to help cities manage rising temperatures. It addresses the challenge of turning satellite thermal data into actionable climate policies for urban areas.

Who Takes Responsibility When AI Agents Make Mistakes

The MIT Center for Information Systems Research says merchants must set clear limits on what AI agents can access, change, or spend. Delegating a task to an AI agent does not shift responsibility away from the human department that owns the process. David Greenberg says every agent needs an owner who handles configuration, access, monitoring, and incident response. Companies must also keep detailed records so they can trace what happened and improve after errors. Teams must translate business strategy into measurable goals and enforceable rules to avoid agents optimizing the wrong metrics.

WeRateDogs Fights AI Content While Keeping Real Puppy Love

WeRateDogs is a popular social media account with about 20 million followers across X, Instagram, YouTube, and TikTok. Founder Matt Nelson has a strict policy of using no AI content in any of their posts or videos. He believes the human connection between people and dogs is special and should not include anything artificial. The company also runs a merchandise business, a podcast, and a charitable foundation that has raised nearly $5 million for needy dogs this year. Nelson started the account in 2015 with a simple dog rating post.

Mural Arts Philadelphia Faces Backlash Over AI Art Banners

Mural Arts Philadelphia unveiled four banner designs by muralist Kelley Prevard that used AI-generated imagery. The designs raised concerns, including a skyline that looked more like Chicago than Philadelphia. Emily Mello, the executive director, did not answer questions about how the AI imagery was approved. The artist has refused to comment. The incident has sparked a debate in Philadelphia's art world about the use of AI in public art and accountability.

NVIDIA FOX Blueprint Guides AI Agents From Defects to Reports

NVIDIA's Factory Operations Blueprint (FOX) is a reference architecture that puts a factory manager agent on top of existing plant systems with specialist agents underneath. At Siemens' Erlangen electronics factory, FOX-based order-planning agents can enable cost savings of up to 20%. The specialist agents include a Visual Inspection Agent using NVIDIA Cosmos, a Root Cause Analysis Agent using Kumo Anomaly, and a Fleet Management Agent using NVIDIA Isaac Sim and ROS 2. Plants can configure which decisions require operator review, giving humans control over critical steps. The system connects operational signals to engineering context like CAD data and simulations.

How to Use AI for Literature Reviews Without Losing Quality

A new guide explains how to use AI tools in literature reviews while keeping research valid. AI can help with searching, summarizing, citation management, and thematic analysis, but it cannot judge research quality. The guide recommends verifying every reference by checking it in a library database or Google Scholar before using it. PRISMA 2020 requires reporting details about automation tools used in selection, data collection, and risk of bias assessment. A joint statement from Cochrane and other groups supports the RAISE framework, which calls for human oversight, transparency, and open reporting of any AI use in evidence synthesis.

Sources

• AI infrastructure grows in Texas with new facilities in Dallas, College Station
• Meet the founder who helped build Nvidia’s early cloud computing infrastructure before leaving the $5 trillion tech giant to bring AI computing power to Africa
• 29% of Employed Seekers Quit AI Interviews [2026]
• PMLR: Machine Learning Research Proceedings Guide
• Machine Learning Framework Advances Urban Heat Island Mapping and Climate Resilience
• Who Owns an AI Agent’s Mistakes?
• How WeRateDogs is fighting AI slop, one perfect puppy at a time
• AI art on Philly’s walls?
• NVIDIA's FOX blueprint sends AI agents from a defect alert to a root-cause report
• Literature Review and AI: Workflow and Reporting