The hottest Safety Substack posts right now

And their main takeaways
Category
Top Technology Topics

Kids and AI

Technically Optimistic • 39 implied HN points • 10 Nov 23
🕹 Technology Safety
  1. Protecting children from online risks should be approached both from a risk-based and a rights-based perspective.
  2. Involving children in conversations about digital spaces, educating them about AI, and designing products with child safety in mind are key steps to safeguarding their well-being in the digital world.
  3. Children's rights in the digital age, including non-discrimination, best interests, survival and development, and respect for their views, need to be prioritized in the design and regulation of technology.

To Pause or Not to Pause, that is the Question.

The Cognitive Revolution • 39 implied HN points • 05 Apr 23
🕹 Technology Safety
  1. The Open Letter from The Future of Life Institute urges AI labs to pause for at least six months in training AI systems more powerful than GPT-4 to develop safety standards.
  2. Different experts have varying perspectives on the necessity and benefits of pausing AI development, with concerns about potential risks of superhuman intelligence.
  3. There is caution and optimism regarding the deployment and potential dangers of scaling GPT-4, emphasizing the need for extreme caution and consideration of risks and benefits.
Get a weekly roundup of the best Substack posts, by hacker news affinity:

The Gordian Knot News "A" List

Gordian Knot News • 168 implied HN points • 25 Nov 23
🕹 Technology Safety
  1. The Gordian Knot News focuses on the importance of cheap nuclear power for humanity's prosperity and environmental conservation.
  2. Understanding key features of radiation and where we went wrong in nuclear power are crucial to solving the issues in the industry.
  3. Proposing a new regulatory system and a radiation harm model are essential steps towards a better future for nuclear power.

A Radical Approach to Handle Your Mistakes

Ruben Ugarte's Growth Needle™ • 19 implied HN points • 23 Jan 24
💼 Business Safety
  1. Volunteering mistakes can lead to safety improvements. In aviation, encouraging staff to report errors has made flying much safer over the years.
  2. Boeing's decline from an innovator to a company facing serious issues shows the consequences of neglecting quality and safety.
  3. The airline industry adopted new practices that prioritized transparency, leading to a safer environment for both passengers and employees.

Can the NRC be Reborn?

Gordian Knot News • 219 implied HN points • 20 May 23
🔬 Science Safety
  1. Reforming the NRC for nuclear power to succeed requires honesty and transparency in communication.
  2. A reformed NRC must establish a compensation scheme for harm caused by radioactive material exposure.
  3. The NRC's focus should shift to emission enforcement rather than reactor design to ensure safety and accountability.

Gnashing My Teeth and Chewing on Painkillers

The Radar • 39 implied HN points • 28 Mar 23
🏥 Health & Wellness Safety
  1. Reflect on past mistakes to learn and improve from them. A moment of stupidity can lead to a lifetime of reflection.
  2. Promote safety consciousness and effective leadership to prevent incidents. Creating a safe environment is crucial for successful leadership.
  3. Design work areas to prioritize safety, minimize risks, and support employees. Safety requires constant evaluation and improvement in work conditions.

Train Derailment

The Polymerist • 249 implied HN points • 21 Feb 23
🔬 Science Safety
  1. Train derailments involving hazardous chemicals occur frequently and pose significant risks to communities.
  2. Enforcement of regulations by agencies like the EPA and DOT is crucial for preventing future incidents.
  3. Long-term solutions to reduce these incidents may require shifting towards safer, more sustainable practices in the chemical industry.

Undoing RLHF, prompt injection, and the brittle limits of moderating LLMs

Democratizing Automation • 126 implied HN points • 18 Oct 23
🕹 Technology Safety
  1. Recent papers challenge the need for safety filters on open LLM weights, suggesting regular releases of parameters.
  2. Fine-tuning LLM safety can be bypassed with minimal supervised examples, raising concerns about robustness.
  3. Moderation in LLMs relates to liability, with Meta emphasizing safety filters in their models, while OpenAI faces challenges due to fine-tuning access.

Deliberate Dysentery

Asimov Press • 96 implied HN points • 21 Jan 24
🏥 Health & Wellness Safety
  1. Human challenge trials have been used for centuries to deliberately infect individuals with pathogens for research purposes, such as testing vaccines and treatments.
  2. The rising popularity of challenge trials can be attributed to ethical reforms, development of solid challenge models, and the ability to safely and effectively test vaccines for various diseases.
  3. Challenge trials have drawbacks, including recruitment challenges, underpayment of participants, and difficulty in quantifying overall impact.

Draft Release Pamphlet for Comments

Gordian Knot News • 95 implied HN points • 16 Jan 24
🏥 Health Politics Safety
  1. The Gordian Knot Group is seeking help from the community for feedback on a release package for radiation emergencies.
  2. The package includes safety items like dosimeters, N95 masks, and potassium-iodide pills, along with informational pamphlets.
  3. Feedback on the draft pamphlet and the need for an illustrator is requested, with incentive for compensation and recognition.

RSV Vaccine Safety Signals Raise Alarms

Who is Robert Malone • 12 implied HN points • 03 Jul 25
🏥 Health Politics Safety
  1. There are serious safety concerns about the RSV vaccines, especially linked to Guillain-Barré Syndrome, which can cause severe health issues.
  2. Some women who got the RSV vaccines during pregnancy faced complications like preterm births, suggesting that the vaccines might be risky for them.
  3. The monitoring system for vaccine safety revealed alarming reports, including deaths and serious allergic reactions, highlighting the need for careful oversight.

Trusted #002 - Finding Common Ground on AI Safety

Trusted • 19 implied HN points • 11 Apr 23
🕹 Technology Safety
  1. AI Safety is recognized as a real problem by many experts in the field.
  2. Stopping AI capabilities development is impractical, as restrictions are unlikely to be effective.
  3. Public support is crucial for government action on AI Safety, as regulations need to address broader concerns like job loss and surveillance.

Brand Safe AI

Addition • 19 implied HN points • 07 Jun 23
🕹 Technology Safety
  1. Brand safety in AI is not a one-size-fits-all concept; it varies based on the specific use case and how AI is implemented.
  2. Design decisions play a crucial role in aligning the level of risk in an AI system with what the organization is willing to accept.
  3. Addressing brand safety creatively involves different approaches like incorporating safety checks, narrow use cases, and extensive testing to mitigate risks.

COVID time series graphs show clearly the COVID vaccine kill people. That's why they keep the plots hidden from view.

Steve Kirsch's newsletter • 9 implied HN points • 11 Jun 25
🏥 Health Politics Safety
  1. Time series graphs can show if a vaccine is safe or not by plotting daily deaths after vaccination. A safe vaccine should show a flat line after the initial period.
  2. Current data for COVID vaccines shows increasing mortality rates after vaccination, which suggests they may not be safe. Many reports don’t show this data.
  3. The medical community often ignores clear signs of vaccine risks, despite evidence appearing in graphs and reports, leading to frustration among those who analyze the data.

Reasons why AI probably won't kill us all

Maximum Truth • 59 implied HN points • 01 Jan 24
🕹 Technology Safety
  1. AI development may not lead to immediate extinction due to limitations in interacting with the real world
  2. Competition between different AIs could help prevent one rogue AI from taking over
  3. Failsafes like targeted attacks against electronics may offer a way to control rogue AI if needed

Manufacturing Jobs Might Come Back: Safer and Fair? Not Clear

Gad’s Newsletter • 38 implied HN points • 08 Jan 24
💼 Business Safety
  1. Manufacturing jobs in the US are seeing an increase in employment due to domestic production efforts and legislation like the Chips Act.
  2. Wage fairness and safety remain concerns in the manufacturing industry, despite some improvements over time.
  3. Enhancing productivity and improving worker conditions must go hand in hand for sustainable progress in the manufacturing sector.

Should we worry?

Sudo Apps • 32 implied HN points • 22 Dec 23
🕹 Technology Safety
  1. AI advancements come with risks like misuse and content flooding.
  2. AI automation may lead to job displacement and increased productivity.
  3. Managing AI advancement involves differing perspectives, safety regulations, and government frameworks.

How to save a life

Force of Infection • 43 implied HN points • 07 Jun 23
🏥 Health & Wellness Safety
  1. Emergencies can happen unexpectedly, so it's important to know how to respond
  2. Watch short videos for quick refreshers on life-saving techniques like CPR and choking rescue
  3. Consider taking an in-person first aid course for more comprehensive training

Why do confounders always work to make the COVID vaccines look unsafe? I asked an expert!

Steve Kirsch's newsletter • 11 implied HN points • 15 Oct 24
🏥 Health Politics Safety
  1. Confounders are factors that can distort data, making vaccines seem unsafe, but they should affect results randomly. It raises questions about why they only appear to show a negative impact on vaccines.
  2. There is a significant difference in mortality rates between different vaccine brands, suggesting there may be deeper issues like manufacturing defects or distribution biases impacting safety results.
  3. Despite individual observations of negative vaccine effects, people are often told to trust aggregated data from authorities, which can lead to doubts about the reliability of personal experiences and observations.

Let's talk about the Olympics

Joshua Gans' Newsletter • 39 implied HN points • 25 Jan 21
🎾 Sports Safety
  1. The costs of proceeding with the Olympics are trivial, and it is safe to go ahead with the event even with minimal attendance in stadiums.
  2. The number of people required to stage the Olympics and be vaccinated is relatively low, making the cost of prioritizing their vaccinations over others very small.
  3. Continuing with the Olympics despite the pandemic provides a sense of normality and value to many countries, with minimal costs and the potential for international cooperation towards global vaccination efforts.

COVID-19 vaccines’ effectiveness and safety exaggerated in clinical trials & observational studies, academics find

Who is Robert Malone • 15 implied HN points • 06 Mar 24
🏥 Health & Wellness Safety
  1. Claims about COVID-19 vaccines' effectiveness and safety were exaggerated in clinical trials and observational studies, impacting risk-benefit analyses.
  2. Observational studies exhibited biases like case-counting window bias, affecting estimates of vaccine effectiveness.
  3. Studies showed issues with safety estimates, short counting windows, and potential negative effectiveness of COVID-19 vaccines.

Faster LLMs, Safer Chains of Thought, and Image Tokenization Reinvented

ppdispatch • 2 implied HN points • 18 Jul 25
🕹 Technology Safety
  1. There's a new book that helps people understand deep learning in a clear way. It covers important topics like neural networks and how they work.
  2. A new technique called Chain-of-Thought Monitorability may help keep AI safe by watching how AI reasons with language. But it’s still seen as a bit weak and needs more work.
  3. Researchers found that recent improvements in AI reasoning might not be genuine. They suggest that better ways to check AI's performance are needed to ensure it really understands and isn't just memorizing data.

Which mRNA vaccine is safer?

Steve Kirsch's newsletter • 5 implied HN points • 10 Jan 25
🏥 Health Politics Safety
  1. The Moderna vaccine might be riskier than the Pfizer vaccine based on some studies, suggesting it has a higher chance of serious side effects.
  2. Recent information indicates that the safety comparison between the two vaccines might not be as clear as previously thought.
  3. Being updated with new data is important for anyone who may help others decide which vaccine to take.

CDC releases V-safe free-text data because they want everyone to know just how unsafe the COVID vaccine really is

Steve Kirsch's newsletter • 13 implied HN points • 19 Feb 24
🏥 Health & Wellness Safety
  1. The CDC has released V-safe free-text comments to show the serious safety problems associated with the COVID vaccine.
  2. There are nearly 400,000 comments in the release, highlighting a variety of negative symptoms and reactions.
  3. Using testimonials from V-safe comments could be a powerful marketing tool to encourage vaccine uptake.

I finally was able to talk to a top official at FDA media relations live. Here's what happened...

Steve Kirsch's newsletter • 12 implied HN points • 15 Mar 24
🏥 Health Politics Safety
  1. The individual made multiple attempts to contact the FDA regarding various concerns, but did not receive any response.
  2. Despite lack of response, the individual was able to connect with a live member of FDA media relations leadership and presented their offer for a public dialog to address differences.
  3. While unsure of FDA's willingness to engage in public discussion, the individual finds the connection made with FDA media relations leadership intriguing.

99 million patient records and they concluded that the benefits outweigh the risks!?!? We respectfully disagree.

Steve Kirsch's newsletter • 12 implied HN points • 22 Feb 24
🏥 Health & Wellness Safety
  1. The study analyzed 99 million vaccinated individuals, claiming the benefits outweigh the risks, but critics disagree due to concerns about safety signals and limitations of the study.
  2. There is skepticism about the benefits of the COVID vaccine, with questions raised about its effectiveness and the reported adverse events, as well as the lack of transparency in sharing data.
  3. The discussion revolves around the comparison of risks between the COVID vaccine and the virus, particularly focusing on adverse events like myocarditis, highlighting conflicting findings in various studies.