I 100% agree, LLMs are a security threat at the moment because and need far more work before I would consider them remotely safe! Users who aren’t technically savvy should not be forced to harbor LLMs on their systems. As the risk of a malicious user breaching and siphoning that data off is ever present. There have to be huge guardrails in place which allow users to have precise control over their data and where it goes.
In regards to iCloud, users should always have a choice as to which apps are opted-in to iCloud at start-up. I know they think iCloud is the best shit, however, letting the user decide is king. The same could be said for all the data harvesting enabled by default on iOS/Mac OS (I vindictively turned that shit off making a WTF face).
As for Apple making Apple Intelligence temporarily opt-in, I’m not sure they would do that. As they’ve seen the outrage caused by LLMs, I think Apple might make an exception and remain opt-in. Though, this is only an opinion and could be proven wrong in the near future.
As for Linux, I did switch almost a week and a half ago to Ubuntu because Microsoft pissed me off! I experienced the pain points caused due to reacquainting myself with the OS, found out several tools I loved and used back in the 16.04 days do not play nicely with 24.04; I borked Ubuntu 3 times before getting it right. ROFL Now it works just fine since Canonical pushed patches that solved underlying issues in their code. I was able to customize and play games, it’s just the lack of proprietary software for iPhone management. I’ll have to get a Mac Mini for that purpose.
The privacy and security issues with LLMs are mitigated by the majority of it being on-device. Anything on device, in my opinion, has zero privacy or security issues. Anything taking place on a server has a potential to be a privacy issue, but Apple seems to be taking extraordinary measures to ensure privacy with their own systems, and ChatGPT, which doesn’t have the same protections, will be strictly opt in separately from Apple’s service. I see this as basically the best of all options, maximizing privacy while retaining more complex functionality.
ChatGPT is a disaster in my opinion, it really soured my opinion on LLMs. Despite your educated opinion on the matter of Apple Intelligence; I have deep-seated mistrust of LLMs. Hopefully, it does turn out fine in the case of Apple’s implementation. I’m hesitant to be as optimistic about it. Once this is out in the wild and has been rigorously tested and prodded like ChatGPT; only then might my opinion on Apple Intelligence be changed.
Is the distrust in the quality of the output? If so, I think the main thing Apple has going for it is that they use many fine tuned models for context constrained tasks. ChatGPT can be arbitrarily prompted and is expected to give good output for everything, sometimes long output. Being able to do that is… hard. However, most of apple’s applications are much, much narrower. Like, the writing assistant which will rephrase at most a few paragraphs: the output is relatively short, and the model has to do exactly one task. Or in Siri: the model has to take a command, and then select one or more intents to call. It’s likely that choosing which intents to call, and what kinds of arguments to provide are handled by separate models optimized for each case. Despite all that, it is very possible that errors can still occur, but there are fewer chances for them to occur. I think part of Apple’s motivation for partnering with OpenAI specifically for certain complex Siri questions, is that this is an area they aren’t comfortable putting Apple branding on due to output quality concerns, and by providing it with a partner, they can pass blame onto the partner. Someday if LLMs are better understood and their output can be better controlled and verified for open ended questions, that’s when Apple might dump OpenAI and advertise their in house replacement as being accurate and reliable in a way ChatGPT isn’t.
I think it’s due to a combination of the tech still being relatively young (it’s made leaps and bounds) and its thoughtless hallucinations that pass as valid answers. If the training data is poisoned by disinformation or misinformation, it makes any output potentially useless at best, at worst it’s harmful. The quality of LLM results purely depends on the people in charge of creating them and the source of its data. After writing it out, I feel that I mistrust the people in control of LLM development because it’s so easy to implement this tech incorrectly and for the people in charge to be completely irresponsible. Since, the techbros behind this latest push for making LLMs into AI are so gung-ho about it, the guard rails have been pushed aside. That makes it all the easier for my fears to become manifest.
Once again, it sounds all well and good what Apple is likely trying to do with their implementation of LLM. However, I can’t help but wonder about how terribly wrong it can all go.
I 100% agree, LLMs are a security threat at the moment because and need far more work before I would consider them remotely safe! Users who aren’t technically savvy should not be forced to harbor LLMs on their systems. As the risk of a malicious user breaching and siphoning that data off is ever present. There have to be huge guardrails in place which allow users to have precise control over their data and where it goes.
In regards to iCloud, users should always have a choice as to which apps are opted-in to iCloud at start-up. I know they think iCloud is the best shit, however, letting the user decide is king. The same could be said for all the data harvesting enabled by default on iOS/Mac OS (I vindictively turned that shit off making a WTF face).
As for Apple making Apple Intelligence temporarily opt-in, I’m not sure they would do that. As they’ve seen the outrage caused by LLMs, I think Apple might make an exception and remain opt-in. Though, this is only an opinion and could be proven wrong in the near future.
As for Linux, I did switch almost a week and a half ago to Ubuntu because Microsoft pissed me off! I experienced the pain points caused due to reacquainting myself with the OS, found out several tools I loved and used back in the 16.04 days do not play nicely with 24.04; I borked Ubuntu 3 times before getting it right. ROFL Now it works just fine since Canonical pushed patches that solved underlying issues in their code. I was able to customize and play games, it’s just the lack of proprietary software for iPhone management. I’ll have to get a Mac Mini for that purpose.
The privacy and security issues with LLMs are mitigated by the majority of it being on-device. Anything on device, in my opinion, has zero privacy or security issues. Anything taking place on a server has a potential to be a privacy issue, but Apple seems to be taking extraordinary measures to ensure privacy with their own systems, and ChatGPT, which doesn’t have the same protections, will be strictly opt in separately from Apple’s service. I see this as basically the best of all options, maximizing privacy while retaining more complex functionality.
ChatGPT is a disaster in my opinion, it really soured my opinion on LLMs. Despite your educated opinion on the matter of Apple Intelligence; I have deep-seated mistrust of LLMs. Hopefully, it does turn out fine in the case of Apple’s implementation. I’m hesitant to be as optimistic about it. Once this is out in the wild and has been rigorously tested and prodded like ChatGPT; only then might my opinion on Apple Intelligence be changed.
Is the distrust in the quality of the output? If so, I think the main thing Apple has going for it is that they use many fine tuned models for context constrained tasks. ChatGPT can be arbitrarily prompted and is expected to give good output for everything, sometimes long output. Being able to do that is… hard. However, most of apple’s applications are much, much narrower. Like, the writing assistant which will rephrase at most a few paragraphs: the output is relatively short, and the model has to do exactly one task. Or in Siri: the model has to take a command, and then select one or more intents to call. It’s likely that choosing which intents to call, and what kinds of arguments to provide are handled by separate models optimized for each case. Despite all that, it is very possible that errors can still occur, but there are fewer chances for them to occur. I think part of Apple’s motivation for partnering with OpenAI specifically for certain complex Siri questions, is that this is an area they aren’t comfortable putting Apple branding on due to output quality concerns, and by providing it with a partner, they can pass blame onto the partner. Someday if LLMs are better understood and their output can be better controlled and verified for open ended questions, that’s when Apple might dump OpenAI and advertise their in house replacement as being accurate and reliable in a way ChatGPT isn’t.
I think it’s due to a combination of the tech still being relatively young (it’s made leaps and bounds) and its thoughtless hallucinations that pass as valid answers. If the training data is poisoned by disinformation or misinformation, it makes any output potentially useless at best, at worst it’s harmful. The quality of LLM results purely depends on the people in charge of creating them and the source of its data. After writing it out, I feel that I mistrust the people in control of LLM development because it’s so easy to implement this tech incorrectly and for the people in charge to be completely irresponsible. Since, the techbros behind this latest push for making LLMs into AI are so gung-ho about it, the guard rails have been pushed aside. That makes it all the easier for my fears to become manifest.
Once again, it sounds all well and good what Apple is likely trying to do with their implementation of LLM. However, I can’t help but wonder about how terribly wrong it can all go.