ArnaulttheGrim

@ArnaulttheGrim@lemmy.world

• ⁨Comment⁩ on ⁨How long do we have before PCs get locked bootloaders and corporations ban installation of "non-approved" software? (for context: Google is restricting sideloading worldwide on Android ETA 2027)⁩ ⁨⁨10⁩ ⁨hours⁩ ago⁩:

  Your statement made me chuckle because I agree about the social fantasies the current models have allowed folks to live.

  I think where I simply dissent is, do I believe corporations are evil and greedy enough to try something like this? Yes, I do. Hell Apple is the epitome of this with all the hoops you have to jump through and any non-basic-tech-literate person would not pursue/approach.

  Because we are simply arguing on an internet forum, Im gonna take my personal belief and assert it here (/s). Looking at the fact that Microsoft backed off the console war, my greater concern is not a ‘gotcha’ that companies like Google and such will apply, its the concept that we are quickly running into fewer and fewer alternatives so its going to be a ‘lesser of evils’ choice.

  Apple showed the way, other companies have been struggling to figure out how they are going to match that. The one foil to this thought/approach is that like the op of this thread stated, an as I infer, users will gravitate away from being locked in.

  Or hopefully we both wind up with wrong conjectures and find that the market goes in a different direction, but I admit myself that is kind of wishful thinking.

• ⁨Comment⁩ on ⁨How long do we have before PCs get locked bootloaders and corporations ban installation of "non-approved" software? (for context: Google is restricting sideloading worldwide on Android ETA 2027)⁩ ⁨⁨12⁩ ⁨hours⁩ ago⁩:

  Call me when they start bringing in legitimacy between understanding how the brain actually functions and relate it to how computation systems actually work.

  The current models STRUGGLE with basic speech comprehension that humans are able to nail with significantly higher precision (Just look at LMLs that struggle with dialects between large regions like the US). Use a slang word in a modern search, or use a common definition vs the literal, AI stumbles and fails frequently. Having worked on models as someone whose job was in AI, the algorithms STRUGGLE even understanding basic concepts such as ‘Yes’ and ‘Yeah’ being interchangeable without dedicated training. There is a reason that it used to be countless humans sitting in a room teaching a machine how to do something with basic boolean values.

  Current Automated Intelligence (I refuse to call it Narrow AI as it diminishes the term AI) will simply be the way of things for a long time until these companies can build trust in them and are able to actually roll out reliable items that: 1) Dont make up data. 2) Can verify data on its own. 3) Can actually understand people when they type/say something.

  True AI or what they are calling AGI nowadays is a pipe dream similar to what 3D/Augmented Reality/Holographic concepts are. There will be spikes of innovation followed by periods of stagnation. The only difference is that right now current AI models are useful in the corporate world which will lead to shorter periods of stagnation comparably.