alkonaut

43 ms·

781.
by alkonaut 2y ago
People seem to always circle back to arguments about whether wfh is better/worse for productivity and so on. But that potential cost/benefit is surely drowned by the extra cost of paying people to work in an office. If productivit
782.
by alkonaut 2y ago
Does IA have much information on users? I’ve been in dozens of these HIBP leaks (including this one) but still none have concerned me, since they were mostly just email/password and nothing else. Does IA store anything sensitive for an
783.
by alkonaut 2y ago
If you write a rust+wgpu renderer now, your options are already Rust, glsl, wgsl (and probably more). You could do Spir-V too on webgpu so long as you stick to desktop. I'm sure we'll see people load glsl on webgpu on the web too.
784.
by alkonaut 2y ago
> The AlphaFold paper has countless authors, many researchers and company resources underlying it. Hassabis’ contribution is management of resources and entrepreneurship, not the actual science. That's usually how you get a Nobel pr
785.
by alkonaut 2y ago
Is this a guest writer? It doesn’t have the Wolfram tone at all. It describes a universe that isn’t centered on Stephen Wolfram, for example.
786.
by alkonaut 2y ago
Yes but that’s where WebGPU really shines though isn’t it? Writing a render graph system (or any reasonably complex renderer that at least does several passes) in Vulkan means juggling with manual synchronization but writing one in WebGPU a
787.
by alkonaut 2y ago
Not at all. OpenGL had that get-something-done quickly feel but it was riddled with limitations and unnecessary complexity. DX12/Vulkan means you do 2000 lines of boilerplate to get anywhere. WebGPU is actually a nice step back towards
788.
by alkonaut 2y ago
Handle text in two ways: either it's controlled by you and you can do simple, efficient, and naive processing, or it's not (it's translated resources, or user input) and you can't. For the former case, you don't nee
789.
by alkonaut 2y ago
This is from 2021 and the main issue the author has with wgsl has long since been fixed. There is a lot less <f32> needed in ~2014~. Edit: in 2024
790.
by alkonaut 2y ago
Well so far it works to just add ”-ai” so it feels like a pretty easy addition
791.
by alkonaut 2y ago
I think the Nobel prize doesn't want any scientific advance to fall outside the range of awards entirely.
792.
by alkonaut 2y ago
Make a search engine that doesn’t have AI result except when I specifically ask for it, or you soon won’t have a search engine business. A really quick fix is to search with “-ai” and that Google doesn’t do this implicitly for images is rea
793.
by alkonaut 2y ago
I feel old. I made my masters thesis on RNN's for learning dynamic systems e.g. for control purposes (quite a novelty at the time, around 2000). We wrote the backprop in C++ and ran it over night. Yes it was slow as hell with the tiny
794.
by alkonaut 2y ago
Source-available is great. It's not open source, but proprietary with source available is still much better than proprietary with closed source. Seeing the source makes it 100x more likely that I understand how a product works, or that
795.
by alkonaut 2y ago
A question that's going to become very real very soon is this: If I video call someone and need them to prove they are human. What do I do? Initially it will be as easy as asking them to stand up and turn around, or describe the headli
796.
by alkonaut 2y ago
I never dared updating my CV because I waited too long and now I just fear it would seem I was on the verge of quitting my job if I suddenly moved even a millimeter on LinkedIn.
797.
by alkonaut 2y ago
If the requirement is having a linkedin page and they don't appreciate this type of humor (they think it's a bad first impression or whatever), then there is no compensation that makes me want to work there. A company that lets a
798.
by alkonaut 2y ago
LinkedIn, on its face was a great idea. A social network for colleagues. It was a Facebook with better spelling and fewer lost neighborhood cats. Then it degenerated into...this. A place where every connection request I get (and refuse) is
799.
by alkonaut 2y ago
I think employment security is one such measure. If employers have to retrain rather than replace, that would help. Some of the value from effectivization is then turned into training. And striking for better job security would make a lot m
800.
by alkonaut 2y ago
Automation creates value. It gets more work done with less. The response to that should be having an argument about how that value is shared in society. You can't stop progress like that, it's a ridiculous stance. Unions should de
801.
by alkonaut 2y ago
Nothing. But that’s not the point. The point is that, to a rounding error, all output is made by a small number of models from a small number of easily regulated companies. It’s never going to be possible to ensure all media is reliably tag
802.
by alkonaut 2y ago
No I don't think technical means will work fully either. But the thing about these regulations is that you can basically cover the 99% case by just thinking about the 5 largest players in the field, be it regulation for social media, A
803.
by alkonaut 2y ago
The immediate danger of large AI models isn't that they'll turn the earth to paperclips it's that we'll create fraud as a service and have a society where nothing can be trusted. I'd be all for a law (however clumsy
804.
by alkonaut 2y ago
If you try to make a business building on top of it it’s a risk you are taking. Whether vscode was partially oss or fully closed seems like an inconsequential detail.
805.
by alkonaut 2y ago
Yes I'm probably biased towards seeing more unsafe as a deal with a lot of wrapper libs (crates which in turn have -sys crates and so on). Looking at the dependency graph, if I use 10 deps directly and 5 have unsafe then that might be
806.
by alkonaut 2y ago
Is there a known ratio of crates that use unsafe to ones that don't? It feels like most nontrivial crates would often need some unsafe. But a system like this might create a scenario where crates offload some of their unsafe code into
807.
by alkonaut 2y ago
> most uses of unsafe would also require explicit whitelisting. I think this is probably where all proposed whitelist/capability proposal discussions end. It's going to be too many crates that are in that category for it to be
808.
by alkonaut 2y ago
That's WAY to much work. I used LastPass, 1Password and Keeper as password managers. None of them is good enough to use that flow. Either the password is short and simple enough that I want to tap it in with a remote (i.e. <8 lowerc
809.
by alkonaut 2y ago
Rarely used passwords is often a "reset password" anyway. But if it's some store or whatever that I use maybe once every 5 years, does it matter what the password is? My point was this: for most accounts, there is no risk inv
810.
by alkonaut 2y ago
A good password - Must be likely to enter correctly on the first attempt, on a bad mobile keyboard, or using a TV remote. - Must be likely to be remembered in my stupid brain even if I haven't used it for many years. Must work even on

More ›