Yeah, that’s fair, but is implicit since I’m arguing against the “sample size is inadequate” POV, not the “there are distributional biases in data” POV. There are a gazillion ways to adjust for these biases (ex. propensity score matching) going beyond just user-base but also including weather type, road type, location, time of day, day of week, traffic density, pedestrian density … that can be done easily with far less than the sample size waymo has. And I bet they do these adjustments.
This! (Thank you for the comment). There’s a reason a 1000 random samples is adequate to reasonably estimate what’s common metrics in a population the size of USA or India (or infinitely large).
They search space for criteria is practically limitless. They have and would absolutely fish for precisely the criteria benefiting Musk.
This playbook has been applied well by the crony capitalist class in the 3rd world, and is always a moving target. Most players know that and will not chase the moving target, knowing that another set of rules will emerge that will create new hurdles protecting the crony capitalist. A few will, and get burned.
There are two reasons to believe this is applicable here:
1. Trump has a track record of quid pro quos (Adelson being a salient
example). Musk is definitely seeking his pound of flesh
2. Lutnick urged people to buy Tesla (shocking and explicit favoritism)
The view that this is just incentivizing local production is naive.
Yeah, there’s no objective and universal barometer for what is or isn’t immoral. I’m providing evidence (as opposed to pulling things out of thin air) for why it’s reasonable for many, or even most, people in the western world to find Guy Rosen immoral. It’s okay for individuals to not find this compelling.
On 2, a few additional quotes from Wikipedia might help (they admittedly don’t directly implicate Guy Rosen, though you’d have to be extremely charitable in assuming he wasn’t party to these decisions):
“ Onavo, which allowed the company to read network traffic on a device prior to its being encrypted, thereby giving the company the ability "to measure detailed in-app activity" and to collect analytics on Snapchat app usage from devices on which Onavo was installed.[12] It did this by creating "fake digital certificates to impersonate trusted Snapchat, YouTube, and Amazon analytics servers to redirect and decrypt secure traffic from those apps for Facebook’s strategic analysis."[13] The program, which was named "Project Ghostbusters" in reference to Snapchat's ghost-shaped logo, was later expanded to include Amazon and YouTube”
“ On January 29, 2019, TechCrunch published a report detailing "Project Atlas"—an internal market research program employed by Facebook since 2016. It invited users between the ages of 13 and 35 to install the Facebook Research app—allegedly a rebranded version of Onavo Protect—on their device, to collect data on their app usage, web browsing history, web search history, location history, personal messages, photos, videos, emails, and Amazon order history. Participants received up to $20 per-month to participate in the program, which was promoted to teenagers via targeted advertising on Instagram and Snapchat. Facebook Research is administered by third-party beta testing services, including Applause and BetaBound, and requires users to install a Facebook root certificate on their phone. On iOS, this is prohibited by Apple's Enterprise Developer License Agreement, as the methods used are intended solely for use by a company's employees (for use cases such as internal software specific to their environment, and internal pre-release versions of apps)”
If everything that anyone cared to not see was censored, there’d be no content on the internet. Also, not smart to conflate (lack of) personalization with government-induced content moderation.
Just want to call out that the head of the trust and safety/integrity division, Guy Rosen, is an Israeli citizen with a strong pro-Israel bias. He’s also a person of questionable morals. From Wikipedia:
“ Guy Rosen and Roi Tiger founded Onavo in 2010. In October 2013, Onavo was acquired by Facebook, which used Onavo's analytics platform to monitor competitors. This influenced Facebook to make various business decisions, including its 2014 acquisition of WhatsApp.
Since the acquisition, Onavo was frequently classified as being spyware, as the VPN was used to monetize application usage data collected within an allegedly privacy-focused environment.”
That Meta considered his questionable ethics a feature not a bug, and repeatedly promoted him, is very problematic.
The language of the corporate meeting room in America is also a lot more casual than in Europe. It has zero bearing on actual professionalism (defined, “ the competence or skill expected of a professional”) given the results of American corporations (or the strong safety record of US pilots). So, while I can literally understand your point, I can’t see why it’s material to anything.
How much of this is intrinsic to nuclear versus due to regulatory requirements (various taxes including subsidies for certain renewables, early decommissioning)? My sense always was that this was about the latter rather than the former, but I’m happy to be educated.
I actually see nothing from Liang suggesting that the work is motivated by or relevant for trading. Care to share your quotes?
What I do see though are these quotes that mildly contradict your POV:
“If we have to find a commercial reason, we probably can’t, because it’s not profitable.”
“We’re going to do AGI. It’s driven by curiosity.”
“Giants have users, but their cash cows shackle them”
“It’s like someone buying a piano for a home—first, they can afford it, and second, such a group of people are eager to play beautiful music on it.”
“Giving back is an honor, and it attracts talent.”
“We won’t go closed-source. We believe that establishing a robust technology ecosystem matters more.”
I personally support trimming bureaucratic fat, but the way the current administration is doing it is the worst way possible - with no due diligence - and will lose public support soon.
Depends on how they weigh the cost of a false positive versus false negative decision. The former seems to often be the key focus of a bureaucracy, slowing down the rate of diffusion of new technologies even among willing adopters.
Bizarre understanding of what happened with Sam Altman and Deepseek. The question to Sam Altman was about whether $10 million would be adequate funding to build a frontier LLM. $6 million is the cost of the _final_ _training_ cost of the DS model. This is an apples to oranges comparison, and the article is just high noise owing to this.
All the references to Jevon’s paradox fail to account for three things:
1. There’s no good forecasting model to account for how aggregate demand moves as a function of efficiency gains in this space
2. Aggregate demand isn’t the same as Nvidia’s share of market, which could drop if alternative paradigms for training or inferencing gain traction
3. Forecasting time horizons matter for discounted cash flow/valuation calculations, which nobody has a good basis for
IMO, there’s just a lot of uncertainty, and it’s fair for the market to discount the optimistic trajectory aggressively based on net new info.