Apple’s business plan isn’t to sell your data to advertisers but rather to sell you (presumably) high quality items that people are willing to pay a premium for.
To Apple, having your data is a liability rather than an asset and thus they’d like to have as little of it as possible yet still being able to offer you the Apple experience.
Task Warrior ( https://taskwarrior.org ) and syncing it to https://inthe.am
Task Warrior is open source ( MIT license https://github.com/GothenburgBitFactory/taskwarrior.git ) as is InThe AM (AGPL https://github.com/coddingtonbear/inthe.am ) and can be spun up without too much difficulty for self hosted options.
New accounts are being created.
How many of those new accounts are people shuffling between different identities and interests? This account is for the music stuff I follow, this is my meme account, this is my serious stuff account…
How much of that is also the people leaving lemmy.world when they were having stability problems?
There’s a pair of videos from People Make Games that digs into it:
curl https://literature.cafe/post/641633
The text ‘A large issue’ appears multiple times in the payload (frankly, surprisingly many).
Inspecting it with xxd:
00004530: 6f6e 652e 295c 6e5c 6e41 206c 6172 6765 one.)\n\nA large
Further down in the output (and the only case of this happening):
000487a0: 6c6f 6e65 2e29 2041 206c 6172 6765 2069 lone.) A large i
In this instance, The A (character 41 in hex) is preceded by 20 which is a space.
The number of times it shows up:
% xxd post.txt| grep 'A large' | wc -l
27
It’s possible that your reader is using the wrong instance of the text… but also a “why are there that many copies?!”
Philosophers haven’t come up with a good way to determine if you were the only conscious being in a universe populated with zombies or not. See also https://en.wikipedia.org/wiki/Philosophical_zombie
I’d also suggest going through https://www.3blue1brown.com/topics/neural-networks to understand some of the concepts behind it. The ‘regurgitating data’ isn’t entirely accurate (and that statement is likely to evoke some debate). The amount of source data is compressed far too much for it to be returned largely intact (see also https://en.wikipedia.org/wiki/Kolmogorov_complexity for some bit on the information theory)… though yes, there are some situations where specific passages (or images) get memorized.
Viewing it on a web page, there are seven paragraphs. https://literature.cafe/post/641633
There are some people who just want to watch the world burn.
A small lemmy site is often has lax administration and no mitigation for the attacks and so it is easy to point a DDOS at it and take it down and then the attackers get to watch people complain about it.
Additionally, the “white noise file uploads” are likely encrypted content rather than “I just want to fill up the database”. I’d be willing to bet that it is CSAM material.
People leaving reddit are barely a blip on most /r’s and the Lemmy’s growth for the past month is less than that of a popular cat sub on Reddit.
Spez would also likely try to avoid doing things that are illegal and would put him or Reddit in legal jeopardy. Random DDOS fall into that category. Spez may be an ass, but he’s more than likely an ass that would rather not waste his time with things that could send him to jail.
AI can be an amazing tool in healthcare, as a double check. For example, assume a doctor thinks you have something. Right now you could have …
Expert systems have been available as part of medical diagnoses for decades. I remember ahem finding one for the Apple ][+ back in the day.
https://pubmed.ncbi.nlm.nih.gov/2663006/ was written back in '89 and you can easily find others going further back.
AI in medical diagnostic capabilities is nothing new or surprising.
I am on an instance that has the finer granularity of the technology topics and the corresponding subscribers to make the posts and comments interesting.
Small, general topic instances means that now I’ve got a dozen copies of the same post scattered through my feed with only a few people on each talking to each other.
Spreading out is ok… but spreading out too far makes people more isolated and content even harder to discover. Why should I subscribe to that instance rather than:
… or any of the next 20 /c that show up that have less than a dozen people subscribed.
I also don’t want to have to wade through a score of posts that Bram Moolenaar died in my feed. One or two will do (one in a general technology /c/ that I subscribe to one in !text_editors@programming.dev ).
What is the value proposition of another /c/technology on a small, general interest instance?
It’s ok to do it… but I am only subscribed to one because there’s a sufficiently large subscriber count and they’re active about moderating. If there was another techlemmy instance that was able to develop a sufficient community of people around the /c/ hosted there, I’d be interested… but subscribing to every instance that is spread out is more clutter in my subscribed section than I want to deal with.
ChatGPT works off of a fixed size possible maximum prompt. This was originally about 4000 tokens. A token is about 4 characters or one short word, but its not quite that… https://platform.openai.com/tokenizer
“Tell me a story about a wizard” is 7 tokens. And so ChatGPT generates some text to tell you a story. That story is say, 1000 tokens long. You then ask it to “Tell me more of the story, and make sure you include a dinosaur.” (15 tokens). And you get another 1000 tokens. And repeat this twice more. At this point, the length of the entire chat history is about 4000 tokens.
ChatGPT then internally asks itself to summarize the entire 4000 token history into 500 tokens. Some of the simpler models can do this quite quickly - though they are imperfect. The thing is at point you’ve got 500 tokens which are a summarization of the 4 acts and of the story and the prompts that were used to generate it - but that’s lossy.
As you continue the conversation more and more, the summarizations become more and more lossy and the chat session will “forget” things.
I was not aware there have been leaks.
The big one was when histories (the prompts that other people used) were accidentally made visible to other users.
https://futurism.com/the-byte/chatgpt-bug-chat-histories-email-phone
https://www.theverge.com/2023/3/21/23649806/chatgpt-chat-histories-bug-exposed-disabled-outage
https://openai.com/blog/march-20-chatgpt-outage
Also consider all the ‘ChatGPT extensions’ that people have written for chrome ( https://chrome.google.com/webstore/search/ChatGPT ) and not infrequent occurrence when someone has an extension with a few tens of thousands of users which gets sold and converted into malware or snooping software ( https://www.theregister.com/2023/08/11/chrome_extension_developer_pressure/ ).
I am making sure that if someone else reads it who doesn’t know or wasn’t sure if I used the words properly that the up front clarification would avoid them asking and then waiting a day or so to get a response that clarifies their question. While you and I may be familiar with it, there are also high schoolers who are reading this who are less familiar with accounting terms and their specific meanings.
https://www.axios.com/2023/08/09/disney-earnings-streaming-profits
Disney also managed to narrow its streaming division losses significantly to last quarter to $512 million from $1.06 billion during the same quarter the year prior.
That progress should give investors hope that Disney is on track to meet its stated goal of making its streaming business profitable by 2024.
https://www.businessofapps.com/data/disney-plus-statistics/
Disney Plus generated $7.4 billion revenue in 2022, an 42% increase year-on-year
They’re getting better and have a lot of other revenue that they can use to offset those losses while they figure out how to do streaming more profitably.
All of the following statements can be true simultaneously:
Another promise was that we’d be able to basically have access to every tv show or film, and that evaporated very quickly when all the studios decided they wanted to gatekeep everything so they could charge even more for it, and then now we have them disappearing so many things that have been out there just because they don’t want to pay royalties to anyone involved.
That is one part of it. The studios believed that if they did the distribution they’d be able to have a larger slice of the pie. This isn’t working as well as they thought it would in many cases (Disney is likely the exception if there is one with a very large historical catalog and several profitable franchisees) but didn’t realize the corresponding infrastructure and (customer) support costs that would incur reducing the amount that they make.
Furthermore, the residuals (things that are currently at stake in striking) are based on “having it available” even if no one is watching it. This means that studios (and Netflix and Amazon) are strongly incentivized to remove shows that aren’t getting watched sufficiently for the draw of having them there because they’re paying residuals no matter if its being watched or not.
Having a streaming catalog of 10,000 shows (Netflix has about 13,000 world wide by some accounts with about 5000 available in the US) would mean they’re paying small amounts to everyone even if they’re not getting watched. If the company (e.g. studio) isn’t set up around the infrastructure and support needed for streaming, this can easily mean that those small amounts can become more than the amount that the company is making off of streaming.
Ever notice how Amazon Prime Video rotates in lots of B movies that are available for a few weeks and then disappear again? This is to get people to watch them but minimize paying residuals for having a bunch of movies and TV series that no one watches otherwise.
With the combination of race to the cheapest for pricing (and sharing of accounts), and infrastructure / operational costs it is quite possible to be in the situation where studios are losing money on hosting shows and at the same time paying actors and writers far less than is fair… and the easiest solution that studios have to resolve this is to aggressively pull shows.
https://text.npr.org/2023/03/06/1161382179/hbo-max-disappearing-shows-series-streaming-warner (full site for a 26 minute audio version)
You can move your account somewhere else, but if you’re subscribed to… say… !technology@lemmy.world then even though you’re on another instance, you’d still be trying to fetch and read content from lemmy.world. If lemmy.world were to go down, then the only content you would see on that /c/ would be from other people on sh.itjust.works and never any of the comments from lemmy.ee or lemmy.ca - who would also have similar problems of isolation of comments.
Lemmy.world going down would mean that the tens of thousands of people who subscribe to /c’s hosted on there would be isolated from each other.
Moving user accounts isn’t sufficient. Moving a /c to another instance is a significant undertaking of trying to get people to do it.
https://startrek.website hosts:
That’s it. Five /c/ all the same topic.
https://programming.dev/communities is software development topics only.
DNS is a poor example to use as there are multiple IPs that can map to one name.
% nslookup www.yahoo.com
Server: 192.168.1.1
Address: 192.168.1.1#53
Non-authoritative answer:
www.yahoo.com canonical name = new-fp-shed.wg1.b.yahoo.com.
Name: new-fp-shed.wg1.b.yahoo.com
Address: 74.6.231.20
Name: new-fp-shed.wg1.b.yahoo.com
Address: 74.6.143.25
Name: new-fp-shed.wg1.b.yahoo.com
Address: 74.6.143.26
Name: new-fp-shed.wg1.b.yahoo.com
Address: 74.6.231.21
You might be interested in looking at something like Usenet News and NNTP where there is one and only one group with the name rec.games.roguelike.nethack ( https://groups.google.com/g/rec.games.roguelike.nethack ) and the contents of that group is federated out to every NNTP server that subscribes to it (not all do). Note that this is a different model and every server contains all the data for the groups that they subscribe to and may exchange those posts with any other server that they federate with (this also solves the “what if Lemmy.world goes down” problem where if the sponsoring server goes down, no one’s comments are exchanged outside of the instance they are on).
https://www.eternal-september.org/hierarchies.php?language=en
But this is a fundamentally different model than Lemmy uses.
Providing just a searchable marketplace that then dispatches you to buy from the store where you have to enter your payment information for them? Sure. And I’d use it. But then, that’s what the ‘shopping’ item on Google is.
If you then say “but I’ve got to enter my payment and shipping info with the store again” - is that as much of a problem as you make it out to be? Because they do need it to do the payment processing and shipping. Fortunately my browser knows who I am and so it auto fills much of that information.
The payment processing is a major hurdle. You’re dealing with each company and each company has its relationship with the payment processor. They have different rates depending on how trustworthy they are and how much business they do. The less trust in the system, the more it costs. The smaller the volume, the more it costs again.
The companies that leak credit card data? They’re gonna pay with increased payment processor costs. You may not see it, but behind the scenes you’ve got https://www.commerce.uwo.ca/pdf/PCI-DSS-v4_0.pdf (that’s a 350 page checklist for the standards for handling payment cards).
Assume that we’ve handled payment… logistics. Amazon has their enormous warehouses and automation that they’ve invested in (the robots are originally from Kiva - https://www.fromscratchradio.org/show/mick-mountz ). Having everything in one place and then dispatching works well and saves money. If I buy something from Best Buy and something else from Pottery Barn and something else from Williams Sonoma – they don’t share a warehouse and so they’re each doing their own shipping with whoever they’ve contracted to do shipping.
Amazon again got big enough that they’ve got their own last mile shipping (and since they’ve got coverage with distribution centers it makes it even more efficient). Shipping from DC to DC is cheap - it’s the last mile that has the expenses.
So, its cheaper (and more carbon / energy efficient) for someone to buy from Amazon and get one package from a marketplace where Amazon manages the payment, inventory, and logistics for delivery than it is to have it be managed from multiple vendors each with their own payment, inventory, and logistics.
Presenting the marketplace is the ‘easy’ part of this. Payment, inventory, and logistics are hard and aren’t solved by federation but rather made more complex and worse from the standpoint of the consumer.