Arguing that many humans are stupid or ignorant does not support the idea that an LLM is intelligent. This argument is reductive in that it ignores the many, many diverse signals influencing the part of the brain that controls speech. Comparing a statistical word predictor and the human brain isn’t useful.
I'm arguing that it's natural for intelligent beings to hallucinate/confabulate in the case where ground truth can't be established. Stupidity does not apply to e.g. Isaac Newton or Kepler who were very religious, and any ignorance wasn't due to a fault in their intelligence per se. We as humans make our best guesses for what reality is even in the cases where it can't be grounded, e.g. string theory or M-theory if you want a non-religious example.
Comparing humans to transformers is actually an instance of the phenomenon; we have an incomplete model of "intelligence" and we posit that humans have it but our model is only partially grounded. We assume humans ~100% have intelligence, are unsure of which animals might be intelligent, and are arguing about whether it's even well-typed to talk about transformer/LLM intelligence.
> You might think "something something incentive systems". No. At my big tech job I had the pleasure of interviewing a few programmers who worked for a large healthcare company that engages in regulatory capture. Let me assure you: They. Do. Not Care.
Regarding programmers specifically I can concur, but with a caveat. Devs often care quite a lot about many things, but often one of those things is not doing the job they were hired for. The tedium of building software for businesses, even what we now call "big tech", is universally unappealing and definitely not the reason most devs started tinkering with computers. So they care very little, and it shows in the tech taking over the clerical aspects of every day life.
The realities of the news business are fine, but the realities of the adtech business are not. As a consumer I very much want profiling and targeting to die off.
Were the ads run on the web not built on a separate business that attempts to violate the reasonable sense of privacy the average person expects, and didn’t attempt to warp consumer’s expectations of privacy, I think there would be less objections.
Was any regex documentation unclear on this? Some libraries have modes that change the semantics of ^ and $ but I’ve always found their use to be rather clear. It’s the grouping and look ahead/behind modifiers that I’ve always found hard to understand (at times).
This is a feature that seems so painfully obvious in the abstract that I’d wager most have never read the documentation. I’ve been a regex user since the early 90s and I’ve never thought about this.
Also an unaffiliated, long term, and happy user of Optery.
If nothing else, I’m glad there are more offerings showing up on this space because of the competition this will hopefully generate.
Consumer Reports also has a semi-related offering called “Permission Slip” that is focused on opting out of data sharing with individual companies, e.g. Netflix, Home Depot, etc.
Many data brokers will not permit third party services to remove the data without a signed limited power of attorney. Note that the power of attorney is limited to interactions for submitting removal requests and opt outs.
Isn't it to be expected? I guess that they have to make demands on your behalf to have your data removed. I guess that's optional because they can still work without it is some cases, and ask you on a case-by-case basis for others, but that's extra work for you and for them, so they may not do it, at least not on the lower tier pricing.
Blame data brokers for making such asinine restrictions.
You can also just use the free version to collect a list of brokers your self and manually contact all of them to find out how much of a pain in the ass it is.
I cleared my name from the net using another service that charged by the month. I paid them for three months, when their work clearing my data from about 100+brokers was completed, then cancelled. 2 years later, my name and personal data still remain no longer to be found like it once was before the scrubbing.
That's great to hear, often they do show up again later, which is why it's a longer-term subscription service. OneRep is the provider for the removal functionality of Monitor, incidentally.
I can't help but be a bit miffed that despite ostensibly being a privacy service, optery is still running a bunch of third party scripts on their site, including google...
I'm curious, what's the point of paying for Optery per year? Isn't removing your data be a one time request. Except for supporting new brokers that might appear.
Your point is spot on. Data removal services have an aspect where a ton of value is obtained in the first 1 - 4 months as the majority of profiles are wiped away, and then after that you're sort of in maintenance mode where the service catches profiles as they pop back up, or when new data brokers are added to the system for coverage.
Optery generally has 2 types of customers:
- The first type are those that care a lot about their privacy and the cost of an ongoing subscription is insignificant to them, so they keep the service running on an ongoing basis for the ongoing automated scans and removals and for getting new data brokers they get coverage for immediately as they are added into the system.
- The second type of customer is more price conscious and is basically looking back and forth between their credit card statement and their Optery dashboard each month and then they either pause or cancel the subscription when they feel they're reached a good stopping point. Optery's pause subscription feature is very popular for this type of customer and you can use it to automatically re-start the service in 3, 6, 9 months, etc.
- Another thing to point out is many other services only offer Yearly subscriptions, Optery offers Yearly or Monthly. If you're price conscious, the Monthly is nice because you can turn it on and off, or pause it as you wish.
More detail on the topic of keeping Optery running on an ongoing basis is on the Optery Help Desk here:
This is a great suggestion and we would like to add this. Not because it would provide any revenue lift though, but because it is what some Optery customers have been asking for, e.g. can I have a lower cost subscription that runs every other month, or every three months, etc. Technically, you can do this today by cancelling and re-starting a Monthly subscription at your desired cadence, or pausing and re-starting your subscription periodically, but that requires manual effort. A configurable cadence is definitely on our backlog though.
Also a satisfied Optery user. Been using their service for the past year, from what I can tell, they seem to have the most robust solution in the space.
I think "partnership" seems like too strong a word for what appears to be the simple use of an affiliate program. Why would OneRep know or care about an individual affiliate and the content of their site, as long as their behavior with regards to the affiliate program is above-board?
Affiliate programs have application processes intended to filter out bad actors and mis-alignment with a brand. To use an extreme example, a web site promoting terrorism would typically be rejected. Approving data brokers as affiliate partners for a data broker removal service is viewed by many as questionable. To use an another extreme example, how would you feel about an anti-virus software company that approved as affiliate partners creators and distributors of computer virus programs.
OneRep is the service I used, briefly. I have no Affilliation with them except as past customer. They delivered as promised and the effect has been persistent 2+ years since the time I discontinued the subscription.
beyonddd should really identify themselves as the founder of a competitor. Nothing wrong with posting, but pseudo-anonymously disparaging the competition seems very inappropriate.
Yes - I flagged myself as an Optery founder on my first comment, but as you mentioned the comment was subsequently flagged and hidden from view. It is also made clear here: https://news.ycombinator.com/user?id=beyondd
From my perspective, I'd put it in any comment mentioning Optery or criticizing competitors. People often read one comment; they don't read all your comments and your profile.
It also adds some credibility: You actually know what you're talking about in regard to this kind of service.
not affiliated with Optery but agree conflict of interest, also misleading by onerep and at best deceptive.
take that potential lack of trust together with the several reports online that onerep's us operation is a sham and they are really operating out of eastern europe and sending user data there...seems shady.
begs the question: what does a privacy-respecting org like Mozilla see in onerep and how is it better than what other companies offer?
Discover's service is limited to only a few sites (which is why it's free). And it is not transparent about progress of removals or requirements.
That might not be the most effective way to reduce spam or reduce targeted attacks, because it ignores many hard to remove exposures.
We have a similar price point at Kanary (I'm the founder) and it covers the resources we invest in the cat & mouse game required to escalate and complete removals on a wide variety of sites, not just a handful of easy ones.
Anyone have experience comparing this to Incogni? I’ve been an unaffiliated user for over a year now. While many brokers have replied, many never seem to.
Optery founder here. We did a deep dive comparison between Incogni and Optery (https://www.optery.com/incogni-review/). The biggest takeaway is Incogni, at this time, does not cover many of the most popular people search sites like Whitepages, TruePeopleSearch, Spokeo, RocketReach, ThatsThem, BeenVerified, TruthFinder, InstantCheckmate, and many others. Most Incogni reviews you'll find online are written by their affiliate partners.
beyondd, I've been reading through this thread and your comments about Optery and you got me to sign up for an account on your site vs Mozilla's service so good job. I was even going to pay for your Ultimate plan for a year. But.... you lost me when I got to the profile page. I have a handful of email addresses and a couple of phone numbers. I would want them all to be scanned for. I had previously been using experian's removal service and they allowed for 10 emails and 5 phone numbers.
Your documentation says:
"You can only select one email and one phone number for scans at this time. However, Optery's engineering team is actively working on providing more configuration options such as the ability to run scans on demand for multiple email addresses and phone numbers."
Any comments on when this will be an option? I would want automatic scans on all of my emails and phone numbers. Not very useful for me without this.
The core of Optery's search functionality is "person" centric. Meaning we start with searches by name, city, state, and age to find "you" regardless of which underlying email or phone number the data broker has on record for you. Because in many cases data brokers may have no email or phone on file for you at all (only home address), or they may have a really old phone or email you have forgotten about. When data removal service scans focus only on phone numbers and email addresses, a lot can get missed. Many people search sites are not even queryable by phone or email, and are only queryable by name, city, and state. Optery does search for phones and emails, but you are correct in that it currently limits them to just one each from the customer at this time. We plan to release the scan on demand feature you referenced in the next few months.
That said, Optery recursively searches through data exposed by data brokers to alleviate the need to input numerous old phones and emails by the customer. In PCMag.com review they said this of Optery's recursive phone number search functionality:
"It uses data found in data broker profiles to recursively expand its reach. For example, in my latest testing, I only gave it my current phone number, but it found records associated with an old number that I used for some 25 years."
Thank you for the reply! I suppose that does make sense, though it still doesn't give a warm fuzzy feel separating the functionality. While the average human might only have one email address they use, I'd venture to say people who would want a service such as this would skew more towards having many they use for privacy reasons.
I get what you're saying about how emails aren't the primary means of finding people, but it is a way, and something people often do have more then one of. I'd humbly request you reconsider and try to better incorporate support for automated scans on multiple emails/phones into the main product. For what its worth it looks like Mozilla's product supports 5 based on their docs.
That said, after submitting this comment I'm going to go ahead and sign up for the one year ultimate anyway in hopes that you will reconsider my request if I'm a paid user. :)
Thanks for the follow up! Scans for multiple phones and emails is something we're working on so stay tuned on that, and don't hesitate to contact customer support with any questions along the way!
Also, you mentioned using Experian's data removal service previously. Do you mind me asking how many exposed profiles the Optery scan located that Experian missed?
I'll go one step further. RTO mandates are being used to mask how shockingly bad most manager's written communication and critical thinking skills are. When their ability to "manage by walking around" is taken away you very quickly notice how terrible their writing (and thus thinking) is.
I've experienced this first hand at multiple companies and it's fairly cringe-inducing to hear the excuses made for simple asks like having a written meeting agenda, project specification, or high level feature descriptions. Work breakdown meetings with these types over video calls amount to them trying to tell other people what to write down and ending the meeting as quickly as possible (perhaps the only upside).
Ascension to management appears to be a more extreme, self selecting version of the Peter principle for many. They can't make a living, or their desired income, in a role that has a well defined output and thus search for hiding places within organizations. Remote work puts them back into a role with expected outputs, and that's basically panic-inducing for someone who opted out of such things long ago.
My verdict: too much pride to admit that whatever communication skills they ever had have atrophied in the cauldron of management-level in-office politics. More energy than is required to learn to write coherently will be expended protecting identity and status. Companies will have to die and be replaced by those with more sensible policies and hiring discernment.
> The best optimization is simply shutting things off
This is the way.
A similar idea has been bouncing around in my mind for a while now. An ideal, turnkey system would do the following:
- Execute via Lambda (serverless).
- Support automated startup and shutdown of various AWS resources on a schedule influenced by specially formatted tags.
- Enable resources to be brought back up out of schedule when demand dictates.
- Operate as a TCP/HTTP proxy that can delay clients so that a given service can be started when it is dormant or, even better, the service isn't serverless but you want it to be. This can't work for everything, but perhaps enough things such that the need to run always on services is reduced.
Cloud Custodian [1] can purportedly do some of this, but I've been reluctant to learn yet another YAML-based DSL to use it.
So this is my "make things designed to be always-on serverless instead" project and the work AWS has done to make Java apps function on Lambda keeps me thinking about the potential to take things that 1) have a relatively long startup time and 2) are designed to be long running service loops, and find a way to force them into the serverless execution model.
> Operate as a TCP/HTTP proxy that can delay clients so that a given service can be started when it is dormant or, even better, the service isn't serverless but you want it to be. This can't work for everything, but perhaps enough things such that the need to run always on services is reduced.
My team mostly builds internal stuff and we save tons of $$$ by using Knative + Karpenter, which basically does that on container + EC2 levels.
Everything I've built in AWS is strictly serverless. You can do an incredible amount with a clever DynamoDB pay-per-request setup, S3 and CloudFront. I haven't once felt the need to reach out to EC2 or RDS and I can't imagine building any sort of control plane to spool them up and down for me.
It covers a broader range of topics than some of the other books mentioned here and helps with an understanding of where existing tech comes from and why. Also it covers topics that you don't necessarily need to know if you're only operating in a cloud environment, but that help to understand what is going on in those far off data centers.
The "Fisher-Price'ing" [1] of the user interface isn't necessarily better in my view. It took the already big and colorful user interface elements and essentially made them bigger. When the video finally gets past the installer there isn't much that wasn't already a part of XP beyond the added, slow UI animations (which is definitely not an improvement, just look at most UI research and anecdotal griping on the web for the last 15 years).
Yearning for XP is something I can relate to. It feels like the last good UI version of Windows. The video is only a nostalgia trip though. Actual improvements on XP's UI would be:
- Removing the hierarchy and modes Microsoft kept piling on since 3.1. These are endemic in the shell. Both the Start menu and Windows Explorer are rife with them.
- Organizing the Control Panel such that it's thematically grouped. All settings involving the display (and thus graphics sub-system) are in a single place, all networking is controlled in one place, etc.
- Reducing the number and complexity of mouse interactions. Context menus should only be needed for the most exceptional cases. Right clicking to accomplish a task should be rare.
- Near elimination of dialogs. Users mostly dismiss them without reading anyway.
- Reduce, to the degree possible, the need to understand the file system, and its hierarchy, to find one's data and make basic use of the system. Watching people try to understand files and the file system hierarchy is pretty painful.
- Reduce the need to manage individual windows, and introduce a way to intelligently place windows across programs such that the user can juxtapose the applications they need to copy/paste between to get work done with minimal effort.
These are just what's top of mind, and I don't fault the creator of the video for making a tribute to XP. I'll add that a lot of these complaints apply to macOS, but often to a lesser degree. Perhaps I'm the old man yelling at the cloud here [2], but this stuff has been eating away at me for years and there doesn't seem to exist the will to address any of it.
All of your UI suggestions are bad UI and shockingly similar to all the BS Windows 10-11 force-feed us.
Hierarchy (organization) is better than search. We all know this because whenever hierarchy is replaced with search, search had to reimplement hierarchy to be usable! E.g., grouping and nesting of pinned Win10+ start menu items.
Control Panel did have groups in Win XP. They suck. I always disable it.
Windows 10 removed a ton of context menus. Windows 11 removed more. They both suck. Stop hiding things from me.
You'll notice I didn't recommend search, or actually anything, as a solution to the problems I mentioned. I think these issues need more research and experimentation before we can arrive at better long term solutions. Also I didn't advocate for hiding things. Instead I think we agree with the idea that less things should be hidden, and the majority of the issues I have with Windows (and macOS/iOS) user interface elements is that they don't do a good enough job of surfacing the things I'm trying to find.
People seem very focused on “should they or shouldn’t they, and why”, which somewhat perplexes me given that there is at least one really good reason for doing this: to create Rosetta Stones for very common pieces of software. To really and truly know if Rust should become the new C we have to start seriously exercising it in the places where C reigns. That’s not a bake-off entirely defined by technological superiority, but also of practical applicability, maintainability, and ability to ship pervasively deployed solutions. This project is a great test of Rust.
RIIR isn’t merely a “sprinkle Rust magic because I believe” thing. The tech has clear potential and I personally think a version of it will take hold where C was once assumed. To know for sure we have to ship more software with it, and we need good comparisons. I’m quite thrilled to have projects like this pressing on.
reply