Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
"Apple is concerned that AI tools could leak the company's confidential data"

WOW... if your IQ is under, say, 90, surely you have problems creating prompts NOT having anything project/company-specific. But not for anyone with higher IQ - and that's the VAST MAJORITY of anything non-third-world country.

I too use GPT-4 for programming (it REALLY speeds up everything!) but make absolutely sure no confidential info is left in any kind of code / questions I post as prompts.

Apple shoots themselves in their foot as they're now reducing the effectiveness / speed of their coders.
ChatGPT can analyze the patterns of requests and discern a lot of information.
 
At first it seems like a spiteful decision (eat own dog food by force) but the reasoning actually makes a lot of sense.
They still are going to probably bungle their own AI effort.
I don't often criticize macrumors, but the title of this is misleading. Apple isn't trying to be anti-competitive; they're just trying to protect their assets.
 
Per science fiction, maybe Apple could end up as the last company on Earth that hasn’t been taken over and controlled by AI forces. Everything else being too complicated for even the smartest humans to understand, let alone control.
 
Generic questions, in which only chore tasks are asked from GPT like "write a Swift data class for this and this example JSON" (with the keynames / values properly changed) is of little help to Microsoft, should they really want to spy on Apple.
You're missing the big picture, but that's okay. I'll let you believe ChatGPT can't figure out what the asker is doing.
 
LOL, sometimes one really doesn't know whether you REALLY think this way or just want to make us laugh :D

It's the truth that Siri has lead us into this AI revolution. Apple saw the potential of this technology long ago and has the global reach with its ecosystems to really drive its strategy forward with even more meaningful AI enhancements.
 
Basically, I think ALL companies banning ChatGPT shoot themselves in the foot. They instead should explicitly ask their employees to remove everything "unsafe" from their questions. I do this all the time during programming and even with this overhead I am a LOT more productive than w/o GPT4 as I can make GPT4 do a lot of chore work, while I can focus on the architecture.
 
Enterprise offerings are coming that wall off corporate data from centralised knowledge pools. OpenAI, Microsoft and Google have all made announcements in this regard.
Right, but until that offering is HERE, don't expect companies to allow ChatGPT if they don't have a contract.
 
Hey Siri, what do you think of ChatGPT?
1684499251749.png
 
Basically, I think ALL companies banning ChatGPT shoot themselves in the foot. They instead should explicitly ask their employees to remove everything "unsafe" from their questions. I do this all the time during programming and even with this overhead I am a LOT more productive than w/o GPT4 as I can make GPT4 do a lot of chore work, while I can focus on the architecture.
It's almost impossible for most companies to do this.
 
  • Like
Reactions: TiggrToo
It's almost impossible for most companies to do this.
It all depends on how smart their employees are. If one can't engineer prompts NOT providing any info on what project etc. one is in, one indeed shouldn't use GPT. For proper prompters, it's a BIG help as it can help in a lot of otherwise time-consuming, chore tasks like the above-mentioned JSON data class generation.
 
It all depends on how smart their employees are. If one can't engineer prompts NOT providing any info on what project etc. one is in, one indeed shouldn't use GPT. For proper prompters, it's a BIG help as it can help in a lot of otherwise time-consuming, chore tasks like the above-mentioned JSON data class generation.
I think people underestimate how much can still be inferred from what the employees of a company are doing with ChatGPT in aggregate, even if they don’t include any confidential information in their prompts. Furthermore, mistakes happen, so some amount of leakage is to be expected.
 
Last edited:
I think people underestimate how much can still be inferred from what the employees of a company are doing with ChatGPT overall, even if they don’t include any confidential information in their prompts. Furthermore, mistakes happen, so some amount of leakage is to be expected.
I mean VERY generic, chore tasks like "write me a MacOS app showing a file selector dialog". It saves a lot of dev time but are still very generic. And MS can only know "this company is working on a MacOS app that shows a file selector dialog".

If the users stay asking such generic q's, MS won't know much of what these projects are. For example, "A Mac app with a file selector and JSON support, which uses URLSessions and closures" (all it can gather from prompts like those of me) - not much of a help :)
 
  • Like
Reactions: amartinez1660
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.