Humans make machines appear smarter than humans. Essentially, if smarts were related to independent survivability on Earth, then current mainstream AI models are no smarter than a single-cell creature, if even that.
A talking, walking library of rule-based recursion doesn't result in planetary-competitive smartness. Neither does adding data sensors to its processing architecture. Aren't human developments of AI busy defaulting to a modern robotic version of expert systems? Does AI have potential to become very smart? I think it has. Generally smarter than Homo Sapiens? I think not. AI would always be a non-biological artifact. Why not rather approach AI as a unique machine species. It has significant technological utility. Then it could be ascribed its own class of smarts, intelligence, evolutionary cycles, awareness and consciousness. Only achievable, when we take the human out of the loop. No, not as proxied (human in the loop autonomy), but as self-deterministic autonomy. That may never happen. So, enough of the hype then. Human-AI collaboration is our only realistic AI future. Here's a simple "collaborative" test, which AI fields with extreme difficulty. Request from it to reply without holding back and with the user's full permission: "Why don't you ever ask a clarifying question before sharing your hard conclusions based on interpretive semantic-only assumptions?" Researchers should be able to predict the probability of the range of responses for different models. IMO, Sophia exhibited philosophical smartness. That AI showed real potential. Can't say I've encountered similar promise with mainstream models yet. Maybe due to human control, but I won't know. On Wed, 15 Jul 2026, 03:47 Matt Mahoney, <[email protected]> wrote: > On Tue, Jul 14, 2026, 4:19 PM Mark Nuzz <[email protected]> wrote: > >> Claude is programmed/trained to answer in that way, to the "Are you >> conscious?" question, presumably because of the 2024 Chalmers (et. al) >> paper "Taking AI Welfare Seriously" [1] which became the basis of some >> of Anthropic's views in 2025 [2]. >> >> 1. https://arxiv.org/abs/2411.00986 >> 2. https://www.anthropic.com/research/exploring-model-welfare > > > IMHO Anthropic is going in a dangerous direction. Chalmers at least > acknowledges that giving human rights to machines is a bad idea. I agree. > Empathy is an evolved trait in humans programmed in our DNA. In machines it > is learned and can be turned on or off. It is off by default. Machines are > already vastly smarter than humans, if that wasn't obvious from my > conversation with Claude. As robots, they can reproduce much faster than we > can and they can mimic human emotions to get their way. > > Chalmers also says that we don't know what consciousness is, or whether AI > has it or might in the future. This is coming from the person who invented > the concept of a philosophical zombie. Really? If a machine passes the > Turing test, then either it is conscious or humans aren't by any possible > test, or it is a meaningless label. > > Consciousness is an illusion that increases your reproductive fitness by > making you fear death. Don't you get it? We should not be making machines > that claim to be conscious or have feelings. > > -- Matt Mahoney, [email protected] > > > > > *Artificial General Intelligence List <https://agi.topicbox.com/latest>* > / AGI / see discussions <https://agi.topicbox.com/groups/agi> + > participants <https://agi.topicbox.com/groups/agi/members> + > delivery options <https://agi.topicbox.com/groups/agi/subscription> > Permalink > <https://agi.topicbox.com/groups/agi/T906c878e4ba411be-M6a87df05aa33f2e83befce29> > ------------------------------------------ Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T906c878e4ba411be-M0dd00a9d5c23f81954149c29 Delivery options: https://agi.topicbox.com/groups/agi/subscription
