Re: Agents and employees may not be equivalent from Amazon's viewpoint
Totally. Agents/software are not the same as employees, but in some cases they can do exactly the same thing. Limited cases now, but they are getting better.
792 publicly visible posts • joined 27 Jan 2012
"Amazon may not like a Perplexity bot because it could jump over some "algorithms" Amazon designed to maximize its profits in goods searches, display and pricing"
100% this. Yesterday I searched for a brand of headphones. Even after using the specific brand filter on top of the search term the first and third lines of results were all sponsored results from brands I didn't want, and the 5th row was a showcase line for a brand I didn't ask for. Literally 50% of the screen was things I didn't ask for.
Forget buying stuff, I'd use an agent that could just list what I actually ask for in a search, but then Amazon loses all that promo money, and if the agent buys stuff it wont buy Prime when it's shoved in it's face and selected by default...
While no doubt return rates are higher can't help thinking this is more about Amazon. In particular:
1 - losing their 'promotion power'. After all a bot wont get distracted by recommendations, constant plugs for Amazon's basics and Prime, and will happily look beyond the first page containing all the sellers that have paid to be there to find exactly the same product for a cheaper price. They also can't tailor prices to different people's perceived purchasing power.
2 - even worse these bots will look at multiple sites so instead of being a site people look to buy things Amazon just becomes another data point for a bot and their brand diminishes. There will probably come a time where people probably wont even visit their site, just ask the agent to find X and after reviewing the options say to go ahead and buy it.
Well I had a customer service agent delete my access to Amazon after leavng a well-deserved 1 star rating for his 'service' (the guy was extremely rude, didn't help at all and left me with a $250 bill for something I returned, Amazon had received according to tracking but I assume then lost).
Didn't lock me out of my account, he deleted it from all Amazon sites. First I knew was when I turned on my Kindle and watched my entire ebook collection vanish before my eyes (luckily I had everything backed up).
Called Amazon the old fashioned way and they told me I didn't exist on any of their ssytems. They had to go to a backup to restore me.
So yeah, what was that you were saying about customer service?
"
But they should understand that the "smart" features are effectively subscription services."
Sorry, that's utter rubbish. Netflix and Spotify are subscription services. The Bose speakers (of which I have none) sonnect to the internet and receive those streams, just like TV's do with regular TV, my phone does with Spotify et al. As far as a customer is concerned they supply the internet connection and the speakers just connect to it like every other device that streams.
Blaming customers for expecting their smart speaker to act like every other internet connected device they own is simply ridiculous.
They don't need the scanner - they'll use free public wifi for wifi scanning or use bluetooth beacons like in Australia to scoop up all the bluetooth IDs. If they have the ID card on the phone they'll ask permissions for the MAC addresses etc and then they'll be able to track everyone (or at least their phones) easily.
It's idiocy if it's mandated for every type of medical condition, but there are many appointments this could work for - eg routine renewal of scripts, receiving blood test results, getting referrals that sort of thing. We use telehealth in Australia for things like that and it's a godsend. Nothing worse than wasting an hour in travel time and the waiting room for a routine 5 minute consult. Also reduces the chances of catching something in the medical centre from all the other sick people.
But totally agree if you have a serious condition that would need a Paramedic telehealth is the wrong option.
"I’ve found quite the opposite. In the domain I know best, it’s consistently knowledgeable, understands context and nuance, and can develop and iterate on ideas in ways that are frankly beyond a simple search engine."
So you work in marketing for an AI firm then obviously.
Non-citizen here and similarly have had no problems. Be polite and respectful and you get a passport stamp and a 'enjoy your stay' in a few minutes.
I'm sure there are some horror stories. I'm also sure there are some horror stories out there with crucial bits left out when they are reported that suddenly seem to make sense when the followups happen...
" the author is American so can't be refused entry at the border"
Actually they can:
"Lawful permanent residents cannot be refused entry unless their travel was not brief (more than 180 days) or they engaged in illegal activity after leaving the United States as defined in 8 U.S.C. 1101(a)(13)" https://www.aclunc.org/our-work/know-your-rights/know-your-rights-us-airports-and-ports-entry
I mean completely agree it's very hard to see how social media posts would qualify as 'illegal' unless they were openly supporting ISIS or something. The ones you hear about getting refused at the border have usually short-term visitors who have posted enough crazy things to get border attention then found to have lied on their visa application, overstayed a previous one or are quite obviously looking to work in the US without a green card.
"Either everyone should be able to offer "Pay or OK" or nobody should."
Nobody should. Simple.
I really hope this sticks. Since Meta pulled this cr@p everybody and his dog has been slapping 'pay or accept tracking from 1096 partners'* on their websites.
*Actually a real number that was on a website today...
The VPN I use (Windscribe) allows you to create a mock location in the system under Android. Works like a charm - if the GPS is turned off Google maps, Uber et al all dutifully report I am in the middle of the city where the VPN server is.
It's the only one I've stumbled across that has this feature but there are probably others that do it too.
(No affiliation with Windscribe, just a user...)
Phones are like computers. A couple of decades ago you couldn't wait to update your computer because the new ones were a huge jump in performance. Nowadays you only update your computer when it dies because the difference between a few years old computer and the new one is negligible.
Phones are the same now - the new ones don't do anything the old ones didn't it's a mature industry.
Hence the reason computer and phone makers are desperately spruiking AI to try and make someone buy their prducts again...
How is it surprising that people are holding on to their phones longer?
The market has long reached maturity - new phones are really just 5% faster than the previous model with a 5% better camera. Plus manufacturers are marketing their phones based on how long they get OS upgrades for, some are promising 7 years.
Under those circumstances the only reason you'd upgrade frequently is because you're a prat that always has to show off that you have the latest phone, or someone who gets their phone free on their phone plan.
If the EU have regulated themselves out of the market whinging about the lack of old phones on the market is not the problem they need to address...
Except the problem is that this will get abused. Google know that the majority of users just click OK, especially when the don't give permission button is reduced to a greyed out text link buried in the permissions box (like they always do with cookie permissions). It also opens the interfect up to a full screen 'approve this or you can't use the website' boxes like the Admiral ones demanding subscriptions.
"good results from asking AI complicated things"
I'd argue you get answers that would please you - not the truth. These things are trained to be yes men that provide believable results you'll be happy with. Sometimes these coincide with the truth, but that is a secondary consideration.
It's the real danger of AI - it's a slightly more sophisticated 'I feel lucky' button from Google search. When was the last time people genuinely thought the first result from a Google search was 100% accurate?
"the scrote is still back on the street nicking phones the same day."
Perhaps you should actually read my comment - it talks specifically about getting them off the streets.
Besides letting 'scrotes' skate with no police intervention achieves nothing.
"Nick the bell end who pays scrotes for stolen phones, and you have a whole bunch of scrotes that have to look elsewhere for their drug money"
Yeah they'll just mug you for your wallet instead, which involves a LOT more risk to the victim. You really need th¡o think things through.
"If you want more cops, then pay more taxes"
I don't want more cops, I want the ones I'm paying for now to actually do their jobs.
"Proven not to work"
Utter garbage. If you take 1 criminal off the street that is 1 less criminal that can potentially break the law and potentially many, many more who will now think twice about it.
Letting criminals run amok and not doing anything about it does nothing to make people safer.
Let's cut through the smokescreen eh?
The solution to not having a global IMEI ban list is to make one.
This is about the UK authorities demanding the ability to kill any phone they choose, just like their demands for backdoors into encryption.
Google and Apple know that if they say they can do it every third-world dictator and authoritarian in the world will be demanding the same power to shut down their opposition.
BTW I have zero skin in this game - running Graphene with zero Google stuff installed so I could care less about their cloud or any bans...
If his move is a negotiating tactic, whereby other countries then come to the table and drop their already-existing tariffs on US goods in exchange for Trump dropping these tariffs then it's better for US businesses selling overseas. If you look at what was happening with Canadian and Mexican tariffs earlier it could be the case.
So the glass-half-full view could be that this is a short-term measure to prompt trade deals where the US had the whip hand. The half empty view being that Trump really thinks the US can ever be cost-competitive with developing countries and import tax his way to prosperity...
The problem with that for browsers is that they have always been free, and there are no new features to add for a browser.
It's the same problem Adobe faced when they inflicted subscriptions on the world - Photoshop had gotten to the point where it did everything people wanted it to do and didn't want to pay for new versions (heck I'm still on CS6 and frankly there's still no new features I'd pay for...) so from a corporate perspective their revenue dried up. S***y deal for users though...
"Privacy costs in inconvenience as well as financially. The de-Googled version of the Pixel Tablet costs rather more than the ad-subsidized version from Google"
Privacy doesn't have to cost. OK I get this is a Murena puff piece, but you can buy a normal Pixel tablet and install GrapheneOS for free.
"Often it will drop something in the summary that I look at and decide "no, that bit is important, I really want that in" and I put it back in"
Serious question - if you have to review things in that depth anyway, and it's only 100 words, is it actually more efficient?
I could do a 100 word summary of a doc I wrote in about 2 minutes and not have to revise it - is the AI really helping all that much if you have to apply so much effort into review? Let's leave aside the nightmarish reality that probably 90% of people wont bother to review output at all...