Architectural Record has posted a fascinating look at some of the reasons behind American housing shortages and high rates of harm to renters.

Why, then, did the U.S. embark upon a path so different from Europe’s? Quite simply, there was a radical divergence in fire-safety approaches during the 20th century and insufficient exchange of best practices. As Stephen Smith, director of the North American Center for Building, argues, American approaches aim to make combustible light wood-frame buildings easier to escape by providing multiple paths of egress; European codes, by contrast, require fire-resistant materials and compartmentation to prevent fires from spreading in the first place. Statistically, the evidence is clear: fire-death rates are consistently lower in Europe than in the U.S. and Canada.

Industry experts in law and AI safety have asked me to write about the fact Tesla lacks general counsel stability.

Apparently this is a canary, suggesting to outside observers that the company engages in illegal practices that no lawyer wants to be around. It is an interesting hypothesis to consider.

Notably, I looked at 2018 news from Reuters.

Tesla Inc has hired Dane Butswinkas, a Washington, D.C. trial lawyer, as its general counsel, the electric car manufacturer said late on Thursday.

Tesla Chief Executive Officer Elon Musk had hired Butswinkas to help settle a case with the U.S. Securities and Exchange Commission, Bloomberg had reported in October.

Musk and Tesla paid $20 million each in fines to settle the case, which stemmed from the billionaire CEO’s Twitter posts in August about taking the company private.

Butswinkas, the chairman of major Washington law firm Williams & Connolly, is replacing Todd Maron at Tesla.

Maron, who also served as Musk’s divorce lawyer, had led Tesla’s legal department since 2013.

Divorce lawyer? Elon Musk used his personal divorce lawyer, Todd Maron, to be Tesla’s general counsel? Kind of makes sense why he was bounced out when the SEC came knocking and investigations got real.

That Reuters report also points to a Tesla PR statement where Butswinkas says he joined for a big and important mission.

Tesla’s mission is bigger than Tesla – one that is critical to the future of our planet.

Sounds important.

Not exactly aligned with reality though. Musk was in deep trouble with fraud allegations and the SEC apparently looked ready to push over a personal divorce lawyer. Different mission, really.

That’s December 2018.

Then, basically after holiday break, when Butswinkas is only starting to warm his seat “critical to the future of our planet” and look around… he quits in January 2019.

Dane Butswinkas Resumes Trial Practice at Williams & Connolly. Williams & Connolly announced today that partner Dane Butswinkas will resume practicing law with the firm this month. Mr. Butswinkas has spent the past several months assisting Elon Musk and Tesla, Inc., first as outside counsel and most recently as General Counsel.

One month. How critical was that mission?

Consider also a total shift in tone when Butswinkas leaves, emphasizing being in a place called home around quality people regardless of any mission.

If home is where the heart is, then Williams & Connolly has always been my home. I’m grateful to the firm for its continuous support, and I now look forward to resuming my trial practice with my colleagues, clients, and long-time friends.

Subtle.

Reads to me like he’s throwing huge shade on Tesla being a shallow, empty and cruel place nobody would ever want to stay.

Then it becomes clear that Tesla can’t hire anyone, as the CEO’s personal divorce lawyer perhaps was covering up things? Or he let the CEO do things nobody else would?

Jonathan Chang, who had worked under the divorce lawyer, was quickly promoted. But by the end of that year, Chang was out too.

Then an Alan Prescott was hired. He lasted almost two years, basically rising to be named Tesla’s top lawyer right before he quits in 2021.

My lawyer colleagues and friends yell at me that this has so many red flags already, there has to be something very bad going on.

No?

General counsel folk don’t abruptly leave a job like this, I’m told. Three in a row, despite the positive narrative of Tesla to the public (e.g. market valuation)?

So then Tesla runs without anyone in the seat during massively ballooning valuations until two years later in 2023 when Brandon Ehrhart is named General Counsel (including a public statement that legal director Dinna Eskin, who journalists suspected of running things, would be pushed aside) to go “hardcore“.

Brandon Ehrhart is company’s sixth legal leader since late 2019

Seventh since early 2019?

I lost count.

That’s a huge turnover rate for the top lawyer job, unheard of and ringing alarm bells, capped with a shift to an aggressive outward prosecution and distraction model after crucial years of no transparency.

I’m not a lawyer, but I can understand why anyone looking at technology safety and the law would have some questions about what might be behind such turnover and new aggressive stance. Who has the answers?

Related: Tesla Whistleblowers Allege Books Cooked Since 2017

Another day, another “veered” Tesla crash, where police wonder why a brand new car suddenly at high speed drives into a fixed object and kills everyone inside.

The Broward Sheriff’s Office is investigating after a driver was found dead inside a burning 2024 Tesla in Broward County on Sunday morning. BSO deputies and Fire Rescue crews found the damaged car on fire near Bonaventure Road and State Road 84 in Weston at around 3:30 a.m. Cameras captured smoke coming from the Tesla’s wreckage Saturday morning. Investigators said the driver of a 2024 Tesla Model 3 was traveling eastbound on State Road 84 approaching Bonaventure Boulevard. As the Tesla entered the U-turn lane to proceed under the Interstate 75 overpass, the driver somehow lost control and drove off the roadway, striking the support column for the I-75 overpass, according to BSO.

This comes after two other recent “veered” Tesla crashes among many prior ones.

In a scenario where an AI system is responsible for delivering real-time updates on the latest advancements in research, drawn from various publications and preprint repositories, several challenges arise due to the dynamic nature of knowledge dissemination. Human experts such as historians, who are trained in such analysis, can carefully assess the validity and relevance of publications, but AI systems must autonomously sift through vast amounts of data, distinguishing valuable information from irrelevant noise and providing coherent updates.

This poses a multifaceted challenge for AI-driven information retrieval and generation. Firstly, coping with the sheer volume and speed of publications requires sophisticated data crawling and indexing mechanisms to ensure comprehensive coverage and timely updates. Secondly, AI models must deeply understand domain-specific concepts and methodologies to accurately interpret and contextualize research findings. Thirdly, confidentiality and integrity must be maintained in compliance with emerging regulations that protect data owners from breaches. Finally, they need mechanisms to track and resolve conflicting information across different sources to ensure the integrity and reliability of the generated insights.

Addressing these challenges involves advancements in natural language processing, knowledge representation, and domain-specific ontology development. Techniques like semantic search, entity linking, and context-aware inference enable AI systems to navigate information intricacies effectively, safely extracting meaningful insights to support informed decision-making in research and development.

Collaborations between AI safety researchers, domain experts, and data scientists are vital for deploying innovative solutions that tackle integrity challenges in AI-driven information retrieval and generation. Leveraging the open standard W3C Solid framework, which offers a distributed approach to data management and access control, is one such solution. By adopting Solid, organizations empower users to maintain control over their data while enabling secure and interoperable data sharing across systems.

Implementing Solid-compatible data repositories and access control mechanisms establishes AI trust frameworks, ensuring data integrity and provenance in intelligence-driven decision-making processes. This allows AI systems to access and analyze data from various sources while meeting stringent privacy and security requirements, reducing risks associated with data exposure, manipulation and misinformation.

Embracing Linked Data and semantic web technologies enhances the discoverability and interoperability of knowledge repositories, facilitating more efficient data retrieval and synthesis by AI systems. Combining AI-driven analytics with Solid’s distributed data management capabilities unlocks opportunities for collaborative research, innovation, and knowledge dissemination, leading to safer transformative advancements in human inquiry and discovery.

In contrast to conventional data storage solutions, Solid offers distinctive advantages, particularly in providing users with granular access control over their owned data (Pods). This nuanced control not only enhances privacy but also promotes high-performance aggregation of data of higher quality and quantity.

Solid Pods diverge from traditional data storage mechanisms by granting users unprecedented sovereignty over their data assets. Through fine-grained access control mechanisms, users can meticulously manage who can access, modify, or share their data, thereby safeguarding privacy and ensuring data integrity.

This heightened level of access control fosters an environment conducive to the aggregation of data so critical to AI success. Users are incentivized to contribute data of greater reliability and relevance, knowing they retain far more meaningful control over its dissemination and utilization than with centralized systems lacking transparency.

Furthermore, Solid Pods empower users to curate data ecosystems tailored to their specific needs and preferences. By enabling seamless interoperability between disparate data sources while upholding rigorous access control standards, Solid Pods facilitate the synthesis of comprehensive datasets and customization of high-efficiency small models, enriched with high-quality information.

In essence, the distinctive feature of access control imbued within Solid Pods not only fortifies data privacy and integrity but also cultivates an ecosystem conducive to the accumulation of superior-quality data assets. This paradigm shift in data storage fundamentally alters the AI landscape of safe information management, fostering richer insights and more informed decision-making.