Invalid pointers and related bugs are older then 20 years…
Languages like Rust don’t solve all problems and nobody is hoping that the research and progress towards better tooling stops there. Even many people working on Rust give plenty of ideas on what future languages can do better or different.
You don’t need to wait 20 years, those discussions are already happening…
So uh… I’m confused on what exactly you are trying to communicate here?
Knowing how your machine works is the only answer. It’s funny you cite the age of pointer issues as some kind of proof. Pointers themselves are abstractions. I’m not saying avoid abstractions or code everything in assembler, but if you want to be a strong practitioner in this field you better be able to disassemble any code and understand what is happening. And alter your practices accordingly.
Hallucinations reduce the success rate of AI workflows, which must be taken seriously.
Imagine a workflow with 8 steps where each step/agent has a 95% success rate, the success rate of this workflow is only (1-0.05)^8 = 0.66 ~= 66%. Not bad but not enough to replace humans yet (unless 66% makes you profitable).
The hallucinations/errors compound and can misguide decisions if you rely too much on AI.
Not enough to replace humans in most critical tasks, but enough to replace Google, that's for sure. My own success rate to find information on Google these days is around 50% by query at best.
I prefer "confabulation," which describes the analogous human behavior where you have no idea what the objective truth actually is, so you just make up something that sounds right
Their developers have intent. That intent is to give the perception of understanding/facts/logic without designing representations of such a thing, and with full knowledge that as a result, it will be routinely wrong in ways that would convey malicious intent if a human did it. I would say they are trained to deceive because if being correct was important, the developers would have taken an entirely different approach.
generating information without regard to the truth is bullshitting, not necessarily malicious intent.
for example, this is bullshit because it’s words with no real thought behind it:
“if being correct was important, the developers would have taken an entirely different approach”
If you are asking a professional high-stakes questions about their expertise in a work context and they are just bullshitting you, it's fair to impugn their motives. Similarly if someone is using their considerable talent to place bullshit artists in positions of liability-free high-stakes decisions.
Your second comment is more flippant than mine, as even AI boosters like Chollet and LeCun have come around to LLMs being tangential to delivering on their dreams, and that's before engaging with formal methods, V&V, and other approaches used in systems that actually value reliability.
Hallucinating has the implication of being wrong. The word further adds the context of being elaborately wrong. That feels pretty accurate to describe an AI going into detail when it is wrong.
Please keep on blogging like you do withoutboats, your articles are a gem that I learn something new from every time.
Due to the work of you and others I do have hope it will all be better in future.
That said, might be my low standards due to many scars from my c++ background, but I’m already plenty happy with what we have today, so the fact that it will get even better in the next years is like cherry on the cake for me.
In defense of the person who wrote the HN title, I’ve seen KYC discussed in front-page articles roughly weekly for the past several years straight. I’ve learned about as much of it as I care to know (and more, honestly) from HN comments on 1st and 2nd page posts in that time. In just the past year, I can see that there have been about 1,000 comments mentioning KYC, and about 21 1st/2nd page posts that are explicitly about KYC (nearly 2 per month). Honestly I don't expect all of HN to know what KYC is, but I did expect most HN readers to have a general idea of what it is and why it's a huge pain for a small % of people (but very large number, 1% of the USA is still >3 million people).
Once you're familiar with it, your brain/eyes key onto "KYC" much more strongly than "know your customer". I might have missed the latter, but "KYC" in the title grabbed my attention instantly and reading the title made my heart jump a bit, because generally KYC means a pain in my ass, and even moreso for friends here on visa.
I have a Canadian friend visiting and staying with my girlfriend and I for a month or so. KYC causes actual headaches for her, to the point that she just decides not to get cellular service at all while she visits unless I get a pre-paid SIM under my name and hand it to her. When she pays for things like restaurants, I can't just Venmo/Paypal/Zelle/ApplePay her back on the spot, I have to withdraw cash at some point and coordinate giving it to her.
The general concept of "KYC" makes sense for some situations, but actual implementations really fucking suck for a lot of people. It's very scary to me to see it be required for more and more categories of services because of the way it's currently implemented.
But remembering the meaning of an acronym while scanning front page post titles without much context? No. My brain is pretty ruthless at evicting TLAs that are reasonably distant from my core areas of interest.
Maybe less important than knowing what it stands for is knowing what the implications are for businesses.
KYC is essentially about knowing who you are doing business with.
For individuals that's relatively easy, just the name and identification is required but typically there is the need to verify that the identification actually belongs to the person signing up. In banking that's why you typically have some video call with a verification provider.
For businesses it gets a lot more complex because it's not enough to know what business your client is, you also have to look through its corporate structure to figure out who the "ultimate beneficial owner" is. Essentially, who is actually controlling the business.
Now it got a lot easier recently as many countries now require businesses to file who their ultimate beneficial owners (UBOs) are.
The painful part is that it introduces friction in customer journeys as now you have to request the documentation.
In the financial industry you also have to run checks on those UBO's so that they are not known terrorists or sanctioned individuals but it seems this regulation is just that IaaS providers need to know who actually operates a server. Presumably for forensic analysis after a cyber attack.
No definitely not, I fully agree with you and others there. Just was a bit surprised by how many of you were there. But that’s okay. Days where we learn are rich days. The richest of them all.
I posted my comment because the linked proposal itself never uses the abbreviation "KYC" and none of the early comments spelled it out, so if (like me) you didn't already know what it means a quick Ctrl-F wouldn't help.
The proposal seems to use the term Customer Identification Program (CIP) instead, mentioning KYC (spelled out) only once, in the introduction:
> Section 1 of E.O. 13984 requires the Secretary to propose, for notice and comment, regulations that mandate that U.S. IaaS providers verify the identity of foreign persons that sign up for or maintain accounts that access or utilize U.S. IaaS providers' IaaS products or services (Accounts or Account)—that is, a know-your-customer program or Customer Identification Program (CIP).
A very significant percentage of us (I suspect a large majority) haven't really bothered with blockchain tech. Blockchain tech doesn't solve any problems that most of us actually need solving.
The use of bone char to filter cane sugar is interesting -- but something that I'm finding puzzling is that the above linked article specifically says that the bone char was used for sugar beet processing, but many other online sources consistently state that bone char is never used for sugar beet processing, only sugar cane processing. Did it used to be used for sugar beet processing but techniques changed?
Based on some reading, beet sugar refining is easier because of less impurities than cane sugar. So the modern technique for producing white beet sugar is to use vacuum evaporation to crystalize out the sugar out of beet juice leaving the impurities behind. Maybe before vacuum evaporation was invented beet sugar had to be undergo chemical processing similar to cane sugar to remove the impurities?
The deeper you plough this field, the more fruitful the findings.
No dead end in sight.
This may even take the cake.
Dark and heavy Belgian cake.
Cake of the dead, so to speak.
Can't tell if this is sarcastic or not but an entertaining example of everyone going straight to the comments (myself included). (for those who didn't see this is link is the same as the main post these comments are under).
