zbentley

58 ms·

721.
by zbentley 1y ago
When I wrote that, I was thinking about typed, compiled languages' documentation generated by the compiler at build time. Assuming that version drift ("D'oh, I was reading the docs for v1.2.3 but running v4.5.6") is user
722.
by zbentley 1y ago
One possible drawback of this kind of system is performance (or broker CPU) getting dragged down by crazy/bad filtering queries. Normally, those issues are solved the usual way (monitor, identify, fix). It’s rarer to see systems that p
723.
by zbentley 1y ago
What an incredibly useful feature. Besides the obvious developer experience benefits, it’s huge for network-bound use cases: really heavily optimized uses of RabbitMQ (or less-optimized uses with really big message payloads) end up bottlene
724.
by zbentley 1y ago
You miss GP’s point. They’re not assuming good faith, they’re pointing out that the government already knows identity credentials and can, encrypted or not, quite easily correlate digital activity with those credentials. The question isn’
725.
by zbentley 1y ago
Sure, but that is as vacuously true as saying “router keeps getting hacked? Just unplug it from the internet.” Huge numbers of businesses want to use AI in the “hey, watch my inbox and send bills to all the vendors who email me” or “get a c
726.
by zbentley 1y ago
Would LLMs help with that? Seems like they could be phished as well. Also, there’s a difference between “know how to be secure” and “actually practice what is known”. You’re right that non-AI security often fails at the latter, but the indu
727.
by zbentley 1y ago
> The fact that there's no foolproof way to distinguish instruction tokens from data tokens is not careless Repeat that over to yourself again, slowly. > it's a fundamental epistemological constraint that human communication
728.
by zbentley 1y ago
The best docs are the ones you can trust are accurate. The second best docs are ones that you can programmatically validate. The worst docs are the ones that can’t be validated without lots of specialized effort. Python’s type hints are in
729.
by zbentley 1y ago
> whether that would still be true had we approached public transit and health care subsidies the same way European countries did? Why wouldn’t it? I’ve heard many different explanations for the US’s wealth, but never that it’s wealthy b
730.
by zbentley 1y ago
Slower? In top speed maybe, but not in time-to-destination (or, given congested streets, average speed). Trains “require” you to make a transfer? Depends on your city, I guess; many train systems are hub-and-spoke-like enough (and dense eno
731.
by zbentley 1y ago
Yes, and? The outcome of that approach is that an important service has uniform low costs to direct consumers, many of whom rely on the service for their quality of life, and many of whom would be unable to afford the service if its costs w
732.
by zbentley 1y ago
> paid through earned income You’re describing a tax on visa holders. That’s an interesting idea; I can think of some benefits and some scary drawbacks/abuses/perverse incentives to doing this as well. Has that been tried anywh
733.
by zbentley 1y ago
Their qualifications for certain roles are easier to vouch for if they have studied at accredited US schools. That’s not to say that all US schools are good or equal , just that the credential is easier to validate.
734.
by zbentley 1y ago
Okay, then show those ads to people who google “midsize SUV” or “AI image generator”, not people who google the specific product by brand name. Sending people who google “midjourney” to competitors’ websites makes as much sense as sending p
735.
by zbentley 1y ago
That would help a tiny amount. The bigger problem, which GP alluded to and which is very, very frustrating to entangle, is the incentives around accountability . Pahlka’s writing puts it better than I could: https://www.eating
736.
by zbentley 1y ago
Citation needed. Other than throughput/reliability risks posed by the revocation check flow (which I know aren’t the reason people don’t use Kerberos on the web, since the big auth providers’ SPOFiness in this area is way worse , as p
737.
by zbentley 1y ago
…yes, that is what I said people need to understand in operator-overloading-happy environments.
738.
by zbentley 1y ago
When it’s Ruby or another language that uses metaprogramming to the point of complete obscurity to tools and humans.
739.
by zbentley 1y ago
C’mon, why don’t you like the authorized secure development workflow? You know, the triple-nested RDP Rube Goldberg machine that connects you to a 15yo desktop environment running on a Pentium III over a dial-up link to the moon (that can o
740.
by zbentley 1y ago
While I don’t agree, I do think there’s merit to the arguments you made (edit remotely, change a line and forget about it). Many anti-IDE arguments boil down to “I like a particular workflow and want to generalize it to all contexts” which
741.
by zbentley 1y ago
Regrettably, “intended purpose” is highly subjective. Sure, << for stream output is pretty unintuitive and silly. But what about pipes for function chaining/composition (many languages overload thus), or overriding call to do e.g
742.
by zbentley 1y ago
I mean, you kinda do. Otherwise you won’t understand why bit-shifting to std::cout prints something, which is pretty much day 1 of C++ hello world introduction (yes, I know there are introductions that don’t use that silly syntax sugar. The
743.
by zbentley 1y ago
That’s an interesting idea. I’m not sure how well hell maps to Banks’ books in terms of choice though. In most religious depictions, one of hell’s predominant characteristics is its inescapability—either logistically or by the mental limi
744.
by zbentley 1y ago
Moderators: this is likely an instance in which the automatic removal of the leading “how” should be un-done, as the HN title looks a bit odd without it.
745.
by zbentley 1y ago
I would hope so as well, but doubt it: if the user consents to their communications being MITM’d by the browser, basically, then I’m not sure there’s currently a legal basis for forbidding that behavior. Many sites/applications accesse
746.
by zbentley 1y ago
This poses a confirmation bias issue though, no? People obviously taking marijuana and sitting around: something you look for because of the stereotype/expectation, and relatively obvious if you’re around them. People doing activities
747.
Google Pixel 10 and C2PA Failures(hackerfactor.com)
2 points by zbentley 1y ago|0 comments
748.
by zbentley 1y ago
Both are very good! I usually default to sharing the Bau Lab's work on this subject rather than Anthropic's because a) it's a little less fraught when sharing with folks who are skeptical of commercial AI companies, and b) be
749.
by zbentley 1y ago
There are also plenty of massive, popular open source projects which don’t require much if any ongoing maintenance. OpenSSL-ish things are the exception, not the rule. Of the rest, it’s fine to keep using old versions of things…however, t
750.
by zbentley 1y ago
I think your assessment of the academic take on AI is wrong. We have a rather thorough understanding of the how/why of the mechanisms of LLMs, even if after training their results sometimes surprise us. Additionally, there is a very la

More ›