In your opinion as an AV safety expert, has Waymo already demonstrated a far higher standard of driving than human drivers in collision avoidance scenarios?
> In your opinion as an AV safety expert, has Waymo already demonstrated a far higher standard of driving than human drivers in collision avoidance scenarios?
That's a difficult question to answer, and the devil really is in the details, as you may have guessed. What I can say that Waymo is, by far, the most prolific publisher of research on AV safety on public roads. (yes, those are my qualifiers...)
Here's their main stash [1] but notably, three papers talk about comparison of Waymo's rider-only (i.e. no safety driver) performance vis-a-vis human driver, at 7.1 million miles [2], 25 million miles [3], 56 million miles [4]. Waymo has also been a big contributor to various AV safety standards as one would expect (FWIW, I was also a contributor to 3 of the standards... the process is sausage-making at its finest, tbh).
I haven't read thru all their papers, but some notable ones talk about the difficulty of comparing AV vs human drivers [5], and various research on characterising uncertainty / risk of collision, comparing AVs to non-impaired, eyes-on human driver [6]
As one may expect, at least one of the challenges is that human-driven collisions are almost always very _lagging indicators_ of safety (i.e. collision happened: lost property, lost limbs, lost lives, etc.)
So, net-net, Waymo still has a VERY LONG WAY to go (obviously) to demonstrate better than human driving behavior, but they are showing that their AVs are better-than-humans on certain high-risk (potential) collisions.
As somebody remarked, the last 1% takes 90% of time/effort. That's where we are...
Very simply, meditation is an attempt at single-pointed concentration. It involves cultivating awareness of the mind's contents and the ability to let thoughts pass without fixation. "Zoning out in the shower" probably means something more like daydreaming, where any and all thoughts are permitted to exist without active control. Focusing intently on a difficult cognitive task ("flow state") is more akin to meditation than zoning out.
A lot of beginners are so bad at this that some amount of guiding back to the goal is helpful. Many can only go a few seconds without getting fixated on passing thoughts.
Practicing one's ability to focus on a single thing and reducing mind-wandering will improve one's capability for concentration.
You're talking about Samatha-vipassanā which is the cultivation of stable attention and mindfulness as two skills. Your skill can be measured by the nine stages of tranquility:
"immersion" as a better translation than "concentration", suggested by Sujato
(can't remember their exact chat about that EBT translation compared to Bodhi or Brahm in whichever of the miriad of Buddhist Society of Western Australia talk/retreat videos I heard it discussed)
mindfulness of body sensation, feeling, thought and principle bringing enough equanimity to start ignoring it all really easy, though the moral aspect can't be separated because doing not wholesome actions will leave you thinking about them
You're saying this as if the result is unsurprising, however it is significant that the performance jumps so dramatically and it is not a fundamental issue of capability, just a bias in the model to be hesitant towards providing false information. That's a good insight, as it can allow further fine-tuning towards getting that balance right, so that careful prompt engineering is no longer necessary to achieve high P/R on this task.
Not at all! I think there's obvious insights being missed by people in how they prompt things. For instance, reality is not dualistic, yet people will prompt dualistically and get shoddy results without realizing their prompting biases are the issue. I see this as evidence AI is calling us toward more intentional language usage.
Did you intend to give your DNA to GSK PLC too? The company that bought the DNA data from 23andme is GSK PLC, the tenth largest pharmaceutical company and #294 on the 2022 Fortune Global 500.
In 2012, GSK pleaded guilty to promotion of drugs for unapproved uses, failure to report safety data and kickbacks to physicians in the United States and agreed to pay a US$3 billion (£1.9bn) settlement. It was the largest health-care fraud case to date in the US and the largest settlement by a drug company.
I intended to give it to anyone 23andme would provide it to. I previously provided it to Harvard’s Personal Genome Project, along with my lifetime medical records. More recently, Northwestern Medicine and a similar program in affiliation with the National Institute of Health (“AllOfUs”).
Yes, it is complex. I agree that providing personal data to trustworthy research programs is beneficial to the public. Do you agree that providing detailed health data to untrustworthy corporations can easily become problematic? Because so far, you've made it sound like you don't see a reason for an individual to not provide their data to 23andme.
It is problematic but has no perfect solution, as there is no such thing as perfect security. Create data security and governance requirements contractually. Require the partner carry insurance as well as attest to and provide evidence of their controls and processes. If they fail to protect the data provided, require penalties outlined in the data processing agreement.
Alternatively, 23andme could offer compute to pharma companies that can run against their genetic data lake, with DLP and data security controls between them and the pharma customer. This would minimize leakage potential while still allowing compute against the data.
Thank you! You mirrored my thought process exactly.
Is there an open genome movement where you can just donate your genome into the public domain? I don't really care who has access to it, but it's a hassle to have to manually apply for each project that wants it.
> I intended to give it to anyone 23andme would provide it to.
So basically any big pharma or big insurance corporation because those are the ones that will get it eventually and will use it for their own profile without regards to any negative consequences for you or anyone else.
I wish the data would be combined with things like Vanderbilt’s BioVU databanks, etc, for actual translational research and not for swift for profit research.
This is just another version of “socialized research privatized profit.”
> The Genetic Information Nondiscrimination Act (GINA) of 2008 protects Americans from discrimination based on their genetic information in both health insurance (Title I) and employment (Title II). Title I amends the Employee Retirement Income Security Act of 1974 (ERISA), the Public Health Service Act (PHSA), and the Internal Revenue Code (IRC), through the Health Insurance Portability and Accountability Act of 1996 (HIPAA), as well as the Social Security Act, to prohibit health insurers from engaging in genetic discrimination. Title II of GINA is implemented by the Equal Employment Opportunity Commission (EEOC) and prevents employers from using genetic information in employment decisions and prevents employers from requesting and requiring genetic information from employees or those applying for jobs.
This is not an objective argument, if you were expecting one based on the title. The author is making this claim as "an antisocial autistic person".
> I can't bring myself to understand the importance for such things (if it even exists.) Communication is about the exchange of information, not superfluous nonsense that contains hardly any relevance to it.
I've played 3000+ games of competitive Halo Infinite over the last year while climbing to top 1% and only had 1-5 games with noticeable cheating. It's nowhere near "ubiquitous and rampant". In the singular game with blatant aimbotters, it was actually a really close and exciting game because while their aim was perfect, their strategy and game sense was atrocious. So with good communication and coordination on our side, we were able to outsmart them and win.
Most people have no idea how to detect basic non-blatant cheats. Even simple low-FoV aimbots sneak past the vast majority of players. Good enough cheats are virtually undetectable by just human inspection.
The few people you saw cheating were probably frustrated enough with the game they decided burn their accounts and ticked enough boxes in their cheat menus to let them go out with a bang. It may of course be that you are particularly good at spotting cheats and Halo Infinite is just blessed to be relatively cheater-free. (I've not played it.) But I'd expect 1-2 orders of magnitude more games to have cheaters than your estimate.
