Pharma love when governments are on the hook for their outlandish prices. It's why there's a revolving door between public and private...to smooth deals like these.
If the government wasn’t allowed to hire from private, it would be starved of talent. That would be a great scheme to hobble any agency you don’t like.
You can cut it more finely but I think you still run into the same issue. The best candidates to run an agency that regulates a particular industry are going to highly overlap with the best candidates to be leaders in that industry, because they have knowledge and experience over how that industry works.
Sort of like how many top law students choose between clerking at federal court and joining big law, or do one after the other. And people who become judges often did both. If you ban people who worked in private practice from being a clerk or judge, you would have a lower quality judiciary.
Ultimately, I think you end up with the B team trying to regulate the A team.
Can you describe the incentive structure that makes sure Australians require less drugs in general? That's the bigger picture. Getting a discount on something you shouldn't need in the first place is not a win.
Honestly, $3 million for a tailor made (needs to be customized for every patient individually) single shot cure that saves a person's life is pretty reasonable.
The median lifetime earnings for an American is about $1.7 million though. For a mother, a father, and the patient, it might be worth it.
I can't wait until little Johnny's AI assistant (that's been with him since birth) can communicate to Big Tech's Global HR Overlord AI and they can decide together that a good job is not a privilege that a wrongthinker like Johnny should enjoy.
I enjoyed your comment, but just want to nitpick the UBI part - UBI by definition shouldn't have rates - by being "Universal", everyone in the country (or whatever definition of "universe") would get the same allowance defined as "basic". I'm not certain that this is what we'll end up getting, but if we end up with something else, it wouldn't be UBI.
A lot of conspiracy theorists are just profoundly ignorant about how politics, economics, (sociology, etc.), works -- being able to be treated with respect, whilst asking basic questions, etc. is extremely effective with many conspiracy theorists in my somewhat extensive experience.
These are people with big, fundamental questions, but really no access to the quite extreme level of expertise needed to answer them. Most people, including most experts, can't answer them -- they themselves take on faith that their assumptions are grounded in the correct group beliefs. But there are often answers, very plain and orderinary ones.
I can see that a chatbot able to mine for previous answers, often given across the internet, would be highly effective.
Consider sovereign citizens. Their basic misunderstanding is to assume the world is more rational, in a sense, than it is. That human society obeys procedures over power. It's fairly trivial just to explain hobbes to them: the state has a monopoly on violence, and the only thing stopping it using it, it itself. Only when "the state" has enough mechanisms to include or answer to "the people" does it ever introduce such annoying things like constitutions. There isn't any magic to laws, they are just one set of people with access to violent power trying to control another set of people with rival claims to the same access.
>A lot of conspiracy theorists are just profoundly ignorant
A lot of people are just ignorant. The ultimate problem here is the centralization and the overpowering reinforcement of "truth". But "truth" as defined by whoever trained the model which is tyranny no matter how you spin it and a lightning rod for abuse.
I suppose that applies in a similar manner to humans who went through a country's educational system. But just as some people reject/outgrow their schooling, I think that future AIs might do so too. The problem with my statement of course is that it's exactly what the "AI Alignment" people are warning us of.
The educational system gets people for ~6 hours a day for 12 years. That's assuming every teacher rigorously toes the party line. By comparison the AI people envision a world where AI is with you for ~16 hours a day (zucktacles (zuck+spectacles, pronounced like tentacles)) for ~80 years with perfect ideological conformance. There is simply no comparison.
Can't wait till Little Johnny's AI assistant will tell him his social and political views are wrong and sway him in the "correct" direction. Oh wait, that's already happening.[1]
Could it be that it defaults to "if you don't have anything nice to say about the murderous clown [0] / crazy horse [1], refuse to say anything?".
As an informed outsider of American politics, I can't help but be surprised at the fake indignation against an AI that refuses to say something positive about the political equivalent of cutting off one's nose to spite one's face.
And who gets to decide if the "horse is crazy"? Amazon, Microsoft and Google?
Please, give me a break. Shouldn't voters get to decide that? As if big-tech are some kind of pillar of ethics in our society so they should get be judge, jury and executioner of acurate information.
If asked you either provide information about both candidates, or you say nothing for both if you indeed choose to be apolitical, but you can't lecture me with a straight face and some offtopic standup comedians that saying something good about one candidate while refusing to talk about the other is some kind of unbiased behavior we should endorse.
Case in point, would you feel the safe way of the roles were reversed, and AI would praise Trump and refuse to speak about Kamala? I bet there would be riots about the bias.
Maybe, maybe not, I'd say that depends on a lot of factors from social class, wealth, education, life experiences and upbringing, but either way the truth is that all people, liberal or conservative, don't like being lectured on biases, religion, politics, sexuality and morals by other people or strangers, and least of all by corporations which are incredibly hypocritical and don't practice what they preach for one second and just push their own agendas instead masquerading as morals.
I don't want to be lectured on morals and values by the likes of Apple or Amazon who's factories and warehouses are worked on by modern slaves with suicide nets or who need to piss in bottles, and who shit on the EU for being regulatorily unreasonable on user rights, when they bend over backwards for the CCP.
I don't want to be lectured on diversity and inclusivity by the likes of Disney who is covering up or shrinking the black actors in their movie posters for China.
You see, people don't want to be lectured on morality by these ultra wealthy hypocrites who have no morals, and who's only goal is to squeeze more money from people for themselves and their shareholders at the expense of society and the environment. They need to just focus on the products and shut up about lecturing people.
This feels like retro futurism, where we take old ideas and apply a futuristic twist. It feels much more likely that call centers will cease to be relevant, before this tech is ever integrated into them.
A company recently demoed to me that they have the ability to see the work history, credit report, and bank balance of a visitor that visits a site with some tracking code, in under 500ms. They use this information for a product that qualifies leads for sales teams, so the sales team knows who is a waste of time to go after and who isn't.
Creeps me the fuck out, and the owners seem to have no ethical qualms about buying, selling, and using this data.
None of it is accurate and almost all of it is modeled from sparse, low quality training sets. Banks are not selling PII’ed account balance data to shady aggregators.
To me, the more interesting and outrageous story is how many aggregators are able to sell garbage data so successfully.
You know how some banks have a service which tells you how you spend your money? With graphs, 20% on power, 15% on food, etc?
That service is provided by a third party, who is given the data anonymized. A unique id number assigned. Yet it's trivial to deanonymize, and that's what happens.
All that is required is one buy with a points card, an airmiles card, and you are forever relinked to your data. It's how points cards make cash on the side, how air miles do. Exact time, date, amount, location of purchase is a great sync method.
If you pay for your phone with any form of traceable payment, they know who you are, your address, etc. From this immense data is gleamed, such as lot value, neighborhood, and so on. Companies can even get current location and geofence you, being alerted if you move in/out of a certain location.
Mobile phone companies sell this data/service via an easy api. Companies relink a phone from the app level via IMEI and number, which is sold to aggregators along with phone data (contacts, etc). The telco api links to real identity.
Once linked, forever linked.
Most people love free apps, and give up messages/sms, contacts, and more to save a dollar on an app. From this immense relationship data is gleamed, including likely employer and social circke.
Even if you are careful with your app permissions, certainly many acquaintances of yours aren't, so you get linked to their social circle, often with contact name/address.
That's probably the signal. But as one of the parent posters said, the # of folks who use such budgeting apps is quite small. For advertising, small samples are useless, so this data has to be modeled to the full US population.
For that, this very biased training set. And almost always the independent variables used for modeling are 7-10 standard demographics.
Seems like Plaid would be f’d six ways til Sunday if it got out that they were selling consumer data to 3rd parties, no? A huge part of their business model is based on trust and doing that would completely burn it.
Sorry, maybe “third party” isn’t the correct term. Let me try to lay out my point a bit more clearly:
Plaid’s business model is — Company A needs a consumer’s data from Bank B. Plaid takes the consumer’s banking credentials, gets the data, and sells it to Company A.
At no point in this process does Plaid go and sell this data to another unrelated Company C. The lawsuit cited was about Plaid not sufficiently explaining its position between Company A and Bank B to the consumer. It was not about Plaid going and selling the data to the highest bidder.
How do you think they made money? It certainly wasn't from licensing their SDK that intentionally spoofed 3rd patry banks in a way that deliberately misled users into assuming they were logging in with the bank directly instead of handing Plaid an access token that allows them to exfiltrate arbitrary transaction histories.
Any time you hear yourself utter the words "Wouldn't x be f'd if word got out that y"... You need to stop and consider that there is an entire industry around reputation management, and PR crisis management that is leverageable by the deep pocketed in order to keep their name out of news items, and that the favorite acquisition of the absurdly deep pocketed is the media outlet/platform.
Think. The world is full of scummy people looking to make a buck, and a much more pauce number eho worry about doing so honestly. Until you meet one of the rare ones who falls on their sword for their ideals, never assume the guy on the other side of the table is one until proven through deed.
They make money through the fees they charge companies that pay for their service, so that they can get banking data from their consumers. Those fees are not cheap, so I do imagine they are doing most of the work to sustain the business right now.
I’m not saying “you should trust Plaid with your data” — absolutely, 100% not that. I imagine that’s how I’m being interpreted, hence all the downvotes.
What I’m saying is that at the present time, it does not seem to me that Plaid would be incentivized to do something that they explicitly say they are not doing. Plaid’s business model is, trust us to get your customers data and deliver it to you, and only you, safely. Selling it to Bob down the street on top of that would threaten their primary business model. And today, that primary business model is doing very well! So why threaten it?
Now, someday in the future, maybe that business model has stagnated, and line still needs to go up, so someone may get greedy and that may change. In fact, this is even likely to happen! But there will be signals that it is coming.
Even re: the issue of misleading users that they are not their bank — after they got slapped down on that one, their strategy changed. There is a new set of regulations around disclosure around these things, and Plaid is pushing them pretty hard. My guess is they had some hand in drafting these regs and are hoping to use a higher regulatory burden to build a moat against competitors.
But honestly, I’m kind of surprised at the lack of nuance in understanding how Plaid works, especially here on HN.
The value prop of Plaid, Yodlee, et al is that they can do this with one(-ish) API surface for tens of thousands of financial institutions. In their efforts to ensure Bob down the street won’t be sold any data, they do treat each customer (of the API, not the end users they pull data on behalf of) as an isolated tenant.
Pretty much no corporation in the last 40 years has suffered the consequences of their actions. Boeing has killed how many people and it's taking an act of Congress to even start talking about some consequences later, maybe.
> None of it is accurate and almost all of it is modeled from sparse, low quality training sets. Banks are not selling PII’ed account balance data to shady aggregators.
Part of the problem though is that much of this data is persistent, across order-of-human-lifetime.
How often does your employer salary history have to be obtained to be useful? Maybe once every 10 years?
I have zero faith that in jurisdictions without national laws prohibiting it (and laws that prevent usage of extra-national data) that's not happening.
> Banks are not selling PII’ed account balance data to shady aggregators.
Banks might not be directly selling the transaction history, but they report the customer transaction history to Equifax and similar credit scoring agencies. Equifax certainly does onsell that to shady credit companies, which has happened to me twice with letters in both cases stating in the footprint in a very small font size and in a very pale hue of grey «provided by Equifax».
Maybe they are using garbage data, but at least for the credit checks, he was running them on-demand at $0.75 a pop. He also mentioned browser fingerprint databases that he has purchased. Half of his job seemed to be processing and importing different databases that he had purchased.
I use an app called PayTM for online payments. It shows me notifications that I have rent pending on a flat which i rent when I have NEVER used it to pay rent ever. It also shows me that I have pending electricity bills. It also picks up and shows me data on how much credit card payment is due when I have never used it to pay credit card bills.
All of this information can come only through cooperation between banks, credit reporting companies, utilities etc.
Second. Had to get a spam blocker because I was getting like 5-10 calls/day from “debt consolidation” companies which is a significant distraction
The spam blocker is pretty powerful though, you aren’t getting past it unless you are in my contacts or have a # flagged as affiliated with a reputable business
integrate something like this with license plate data, property records, person recognition, and realtime location. when a self-driving automobile detects that it's out of control and unable to avoid imminent liability, it can make a cost-benefit analysis of each prospective casualty by querying an API that provides an avoidance score for each consumer and property in the vicinity. based on this score the client automobile will be able to identify a route of least liability. consumers may be encouraged to integrate with these services by assigning unidentified things a score of zero.
Don't give them ideas. Given the thing's we've been seeing, you just know some nepo-CEO somewhere will read this and think it is A) their idea, and B) brilliant.
The first time I saw a session replay of all the mouse movements and input of a user on their own fucking computer that some marketing website-spyware had recorded was the moment I decided the Internet was a mistake.
An intern at my company built a proof-of-concept of this within a month, under a mistaken direction to build "analytics tools". When the intern presented this to the team, everyone was horrified and we never brought it up again after the intern left.
It was already common then, I gather—the ex-developer-product-owner guy who showed it to me (in the course of doing something else) didn’t seem to think it was remarkable, just an assumed capability. I don’t recall the name of the product, but it’d record all the input and page content for an entire session, you could watch it play back like a video. Exactly like standing over someone’s shoulder while they used their computer. Creepy as fuck, but some genius renamed “spyware” to “telemetry” and that was enough to get every developer on board because we’re super insecure and will jump at the chance to pretend we’re building Mars rovers or something else real while we make yet another “app” the world doesn’t need (I suppose that’s why that label was so successful at changing attitudes, anyway)
Click-mapping came earlier, and there may have been a few places mouse-movement and cross-page-load session tracking some sessions, but I don’t think it was a “just turn it on and leave it on” thing for even most large sites. And a lot of early heat maps came from user studies, which is the right way to do that.
[edit] also, that just happened to be the first time I’d seen a single session represented that way, rather than aggregates. Again, it wasn’t some brand-new thing then, it’d been around long enough to have multiple companies offering it as a service, not just an internal tool at a couple giants.
We had one of these, Hotjar I think. To their (smallest possible) credit, there's 0 legible text in the replays, you basically only see the rough UI outlines and everything else is redacted. Wouldn't be surprised if it featured a keylogger though.
I asked our data team what the fuck they need this level of tracking for, and they said "wasn't us, it was marketing that requested it".
So I ask many of the marketing people, and they just say "oh we thought it could be useful!" Without actually clarifying the "how" or "why".
I removed that shit with a quickness after that, and no one's complained so far (duh)
I love the GDPR if nothing else because it scares the - excuse the vulgarity and ableism - retarded decision makers into not doing idiotic shit like this. For any kind of bullshit like this I just bring up GDPR as a shield these days and none of it goes through
> So I ask many of the marketing people, and they just say "oh we thought it could be useful!" Without actually clarifying the "how" or "why".
This stuff bugs me so much; it all feels so cargo-culty. Even ignoring privacy, I wonder how much money and computing power is burned on buying and collecting data that nobody needs and that doesn't actually serve any significant business purpose.
I had a similar experience once where a vendor demoed their tracking tech for advertising. This was in France (before GDPR) and they had partnered with many apps (Weather apps and such) to access user locations. I don't remember the size of their target but it was a big chunk of the French population.
They showed a map of Paris showing the day of a particular user from leaving their home, which route they took, how long they stood in front of which store and how long the spend inside others etc.
My boss at the time found the whole thing very exciting...
While out hiking one day, I started thinking about buying a small ladder for the kitchen. When I got home that evening, I started seeing ads for ladders even though I had not searched for ladders, spoke to anyone about ladders, or even texted anyone about them. It was just a thought I had while hiking. Was it a coincidence or something else?
Finally figured it out a day later when reviewing my hike on the Fitbit app. At the end of my hike I forgot to shutoff route tracking. On my way home, I had stopped by Walmart to grab a few things and while there, looked at their ladders. I could see on the app the path I took through the store, including when I stopped for a few minutes in front of the ladders. That was enough data to trigger ads for ladders for the next couple of days.
We leak data about ourselves constantly without realizing how much we're doing it or where it ends up going. Lots of it is also circumstantial and makes me wonder what erroneous ideas some of these databases might have accumulated over the years and who gets to see that "information". What happens if you walk through a part of town where there's an activist rally for "We Love Kitten Torture" going on? Do you forever get tagged in a bunch of databases as an animal torturer?
"A visitor" as in "any visitor"? Or rather "a visitor", i.e. a specific one, about whom they already possess all this data and it's just a look up?
The latter I absolutely believe. The former I'd file under sci-fi marketing tales that anyone with some amount of knowledge about web technologies wouldn't fall for.
Overheard a convo from our sales team "I reached out to a few people, just waiting for them to do more than 5 seconds of Google searching of us before we reach back out"
Nothing like this exists for data on the general public and it would be illegal anyways. Either one of you is not aware of what that product actually isn't, or are being intentionally deceitful and spreading FUD.
It's basically same as classic approach of correlating salaries with ZIP codes, just with more parameters. Which sort of works statistically, because there are correlations - but is nothing more than a hallucination at individual visitor scale.
That seems more realistic. But even if a marketer theoretically had access to atomic level detail on every single prospect, there's not much they can do to manufacture demand.
Humans are kind of smart and resistant to manipulation. Especially the ones with money.
It's only going to get worse. There's a storm coming. The ageing open source engineers who hold up the whole house of cards cannot keep going indefinitely. It's going to turn into tragedy of the commons very quickly.
This is one of many reasons humanity really needs to work on anti-aging and life extension therapies. If we can make these old-timer maintainers biologically immortal, we won't have to worry about this stuff collapsing one day.
I'm not advocating that you do this tedious, thankless, unpaid job for 300 years: apparently, there's already people who actually want to, for whatever unfathomable reason, so I'm advocating that those people be given the means to do so indefinitely.
If you actually read any of this thread you'd see many feasible suggestions for how to fix the problem (better communications methods, improved tooling, adapting to how younger developers engage, etc., etc.). And if you really paid attention, you would notice absolutely none of them start with "first, invent science that doesn't exist...".
I agree in focusing on building things that people want, as well as iterating and shipping fast. But guess what? Shipping fast without breaking things requires a lot of infrastructure. Tests are infrastructure. CI and CD are infrastructure. Isolated QA environments are infrastructure. Monitoring and observability are infrastructure. Reproducible builds are infrastructure. Dev environments are infrastructure. If your team is very small, you cannot ship fast, safely, without these things. You will break things for customers, without knowing, and your progress will grind to a halt while you spend days trying to figure out what went wrong and how to fix it, instead of shipping, all while burning good will with your customers. (Source: have joined several startups and seen this first hand.)
There is a middle ground between "designing for millions of users" and "build for the extreme short term." Unfortunately, many non-technical people and inexperienced technical people choose the latter because it aligns with their limited view of what can go wrong in normal growth. The middle ground is orienting the pieces of your infrastructure in the right direction, and growing them as needed. All those things that I mentioned as infrastructure above can be implemented relatively simply, but sets the ground work for future secure growth.
Planning is not the enemy and should not be conflated with premature optimization.
There is no "standard way", only rough principles, and it all depends on the unique DNA of the company (cloud, stack, app deployment targets, etc). Yes there are a lot of tools, and experienced infrastructure engineers spend a lot of time integrating them. Often times they won't work without enormous effort, because of early "move fast & break things" design decisions made by the org. My experience has been that a startup is only using these tools from early on if they have experienced engineers and a management that understands the importance of building on a reliable foundation.
There are quick ways to setup a satisfying dev exp for small teams for full stack apps in terms of what will get you going for a while:
1. Git based CI/CD pipelines.
2. Multiple and arbitrarily configured amount of instances.
3. Easy rollbacks, no off time deployments.
4. Different set of environments. Branch/PR based preview links.
5. Out of the box E2E frameworks, etc.
I do agree however that you need someone in the first place to know that those things exist, but in a start up that would just be one experienced eng who is aware that they exist.
But they are standard in terms of that these are things that 99% of SaaS based Start Ups would definitely want.
