s from newcomers about AI and Co-Pilot?
s from newcomers about AI and Co-Pilot?
I thought I had a good grasp of AI until the new Co-Pilot PCs were released. Before that, I saw AI as a kind of server-based system where you send a question or task to a remote AI in the cloud and receive a response, artwork, completed homework, etc. Now it feels like AI is embedded right on your computer. What’s happening here?
First, the Co-Pilot PC name sounds a bit mixed up since I’ve mostly seen Co-Pilot laptops. Are there any desktop versions available? Why is it only offered for laptops if that’s true? Then I learned that Co-Pilot processors or NPUs are compatible only with Snapdragon chips, but Intel is launching one at the end of the year. Meanwhile, some big brands like Lenovo claim you can buy an Intel Co-Pilot laptop right now—so what’s going on? Did they just shift their timeline?
I also wondered what Co-Pilot actually does. I heard about the recall feature, which seems the most interesting to me. I doubt I’ll need it. But there are other AI features too. I worry this will be similar to Amazon Alexa—so many specialized functions that you might not realize what AI can or can’t do, which becomes really annoying when you try to figure it out. What are the main uses for Co-Pilot?
Also, it appears you need at least 40 TOPs to use Co-Pilot. Is this different from system-wide TOPs, which you might have many? It seems Microsoft only focuses on NPU TOPs for Co-Pilot. Why the difference? And how can I check if my old PC has any TOPs? Also, it looks like RTX40XX cards have some system TOPs, but what about RTX30XX? Is that outdated?
Finally, how many TOPs are ideal for a new computer or laptop? The more, the better, or is 40–45 enough? And what does having more TOPs actually give you? Thinking from a regular user’s perspective—someone who uses their PC for gaming, work, media, blogging, video editing, coding, and network tasks—what would be the best approach?
Please help clarify this, thanks! (I searched online, but it only made things more confusing.)
It appears AI is operating within your system. Isn't that strange? What's happening here?
My skeptical perspective suggests AI aims to control everything properly or improperly, as long as it generates clicks and gains profit.
I think this might be a valid point.
These are just my reflections on the topic.
This isn't meant to be a fully reliable statement, it includes many subjective views and overlooks several points.
The system described is essentially a sophisticated algorithm built to absorb data and replicate human interaction. It can only acquire knowledge from its training, making it unable to generate original concepts or insights. Its processing demands are high, favoring modern GPU architectures with NPU (ARM) components instead of older Intel/AMD x86/x64 setups running on Windows 11.
The requirements for Windows CoPilot+ build upon those of Windows 11, which have been available since 2018. However, the integration of CoPilot+ into ARM-based systems is still emerging, especially in 2024, creating variations in software behavior and compatibility issues.
Whether you're using an ARM processor or a traditional Intel processor, and whether applications are running on emulated x86/x64 environments, there are notable differences. A major challenge remains persuading users that this setup offers a meaningful upgrade over Windows 10, which has struggled to gain traction so far.
I'm not an expert, but generative AI has the ability to produce fresh data and models by examining patterns in its training information. It identifies trends and probabilities within the data and builds new outputs based on those insights. You're correct that generative AI doesn't think like humans, and it can't make sudden intuitive jumps, yet it learns by transforming existing knowledge into something novel that fits what it has seen before. This is why I believe it might be applied in products such as Windows—offering users a smart set of alternatives that align with their unique actions but share similarities with common choices.
Because generative AI cannot produce truly original content from scratch, it relies on existing data inputs to operate within a machine learning framework. It lacks an innate sense of morality, truthfulness, or humor, which is important to consider given its dependence on internet-based data sources. In situations where data is scarce or absent, the system often generates information by combining unrelated pieces or copying large volumes of text. A growing concern is the shortage of relevant data for learning. It appears there is a significant gap in understanding how the human mind functions in everyday activities, especially when attempting to replicate cognitive processes in robots. One challenge discovered during efforts to emulate human brain function in robotics is that we typically focus most on filtering distractions, which can impair our ability to concentrate on important tasks—similar to symptoms seen in schizophrenia. If all stimuli are perceived as equally significant, rational decision-making becomes difficult, potentially leading to irrational behavior.
From Forbes magazine, "Although GenAI systems demonstrate remarkable performance, their capacity to address complex, multi-faceted societal challenges remains limited. They perform well on specific, narrow tasks but fall short in developing the broader comprehension required for strategic decisions or ethical considerations." May 9, 2024