austinjp

46 ms·

481.
by austinjp 4y ago
It would be more reassuring for you to explain exactly how you do use the data. Are you training neural networks, even with aggregated/anonymised data? Are you creating synthetic data? I'd be specifically interested in exactly
482.
by austinjp 4y ago
> Re the affiliations, it is correct that these institutions aren’t directly endorsing iollo. You may want to talk to the relevant departments in the universities you list, their "innovations" or "spin-out" teams. Typ
483.
by austinjp 4y ago
Define "successful".
484.
by austinjp 4y ago
Prediction, operating at the limits of human ability in an extremely limited (but very impressive) application.
485.
by austinjp 4y ago
That.... seems like the worst approach?
486.
by austinjp 4y ago
Testers, surely.
487.
by austinjp 4y ago
I think you're somewhat misinterpreting his point. He's not saying "I'm correct whether you like it or not". Instead, he's saying "we are all prone to cognitive errors, whether you like it or not". Ad
488.
by austinjp 4y ago
Any application used within any organisation will eventually be used to communicate notes between users. Shared calendars get polluted with non-appointments like "Hi Jane, thanks for referring Bob, but please be aware of blah protocol
489.
by austinjp 4y ago
It takes effort to understand this stuff. Don't worry: that's normal. If you want to understand it, treat it like anything else you want to learn. Make notes, review them, and prepare before the meetings. A simple way to begin is
490.
by austinjp 4y ago
Why? * * By which I mean with references. And yes, big farma, etc etc, but really, decent studies please.
491.
by austinjp 4y ago
Ah yes, thanks!
492.
by austinjp 4y ago
The daftness of Kapoor's exclusive right to Vantablack prompted artist Stuart Semple to create Black 2.0 and 3.0 (as mentioned by parent) with a license granting use by anyone except Kapoor. https://stuartsemple.com/p
493.
by austinjp 4y ago
Aah come on. Nobody said the Democrats are "adjacent" to stoning women. Sheesh. From the perspective of a country that is to the left of the Democrats, yes, they appear centre-right. I'm not sure you've can offer a truly
494.
by austinjp 4y ago
Your GP "yelled" at you and "made up" random diagnoses to save a few £ per month, deliberately putting your partner's health at risk? Any GP doing what you say would lose their job and not be allowed to practice aga
495.
by austinjp 4y ago
Not sure I understand your API example, but fwiw in Python I use xmltodict [0] specifically the unparse function to convert a dict to XML. I combine this with ujson for speedy JSON [1] and defusedXML [2] for safety. [0] https://
496.
by austinjp 4y ago
> a paperclip factory isn't self-reproducing (that would be a paperclip factory factory) This seems an odd refutation for several reasons. First, the paperclip AI might determine that self-reproducing factories would be an optimisat
497.
by austinjp 4y ago
> My intuition is that there are an infinite number of ways to fail, and the number of ways to succeed are much more scarce This may well be true, but the factors for success are hyper-sensitive to the context of the time with a hefty da
498.
by austinjp 4y ago
I swear, I have zero understanding of why Mac users are so delighted when they discover previously-unknown features. These features should be easily discoverable . But maybe I'm missing a trick. Maybe there's a deliberate effor
499.
by austinjp 4y ago
> The inability to have an honest and open discussion about the future of the NHS means it’s just going to continue lumbering on until it collapses Agreed. Unfortunately, it often seems that "honest and open discussion" rapidly
500.
by austinjp 4y ago
> Everything goes through GPs, that are the medicine homologues of a guy that works in IT and can set the background of his Drupal blog. I don't follow you. GPs go through specialist training and assessment (general practice is, per
501.
by austinjp 4y ago
> Elizabeth Line purple (Pantone 265) has been chosen primarily to assist passengers with wayfinding and navigation. It is visually distinctive from the other modal roundels – London Underground red or London Overground orange – enabling
502.
by austinjp 4y ago
> asking [neural networks] to explain themselves using their own capabilities Exactly. This could be profound. I'm looking forward to further work here. Sure, the examples here are daft, but developing this approach could be like un
503.
by austinjp 4y ago
Surely: training dataset of "personal information" contained genome-related images.
504.
by austinjp 4y ago
Data still has to get to the cache, get stored, and sent to clients. More data = more energy, cache or no. Your point about video streaming being a far larger amount of data is of course fully valid. But I was replying to a thread about Jav
505.
by austinjp 4y ago
Fully agreed. Two aspects that I don't see acknowledged often enough: (1) Not everyone has fast internet. This is particularly the case in rural locations and developing nations. We're witnessing internet inequality and heavy webs
506.
by austinjp 4y ago
Sounds like an infinite procedurally generated VR game of Red Dwarf. But not this https://reddwarf.fandom.com/wiki/Red_Dwarf_XI:_The_Game
507.
by austinjp 4y ago
I still don't see the prerogative for bitcoin. Yes, most industrial processes are inefficient and we can and should reclaim lost energy. There are many ways to do this, and many purposes. Charging batteries seems the most reasonable: m
508.
by austinjp 4y ago
BT is shit. The only consolation: it seems to be equally shit no matter the device/OS/etc provider. With one exception, a "JBL Go" portable speaker I bought years ago. Rock solid, I take it on every holiday. But otherwis
509.
by austinjp 4y ago
The pie totals 200%.
510.
by austinjp 4y ago
Can you explain a little more about how and why this was practical?

More ›