|
|
disable_google_ai_on_chrome
|
|
|
epitome of incomprehensibility
|
In case anyone finds this useful...it worked for me, at least temporarily. Instructions based on Elle Logs' video https://www.youtube.com/watch?v=l4qKqHymCBM; I also posted a similar thing on red blather. 1) Click on three dots on top right hand corner of screen 2) Click "Settings" near bottom of menu 3) Search "manage search settings and sight search" 4) Once there, scroll to "site search", click on it, and fill in the blanks in the pop-up window; it doesn't matter what the first 2 entries are, but the first functions as a name so you can put in something like "non-ai search" 5) *Important*: fill in the third blank with https://google.com/search?q=%s&udm=14 6) Click on three dots next to your new entry in "site search" (e.g., "non-ai search" if you called it that) and select "Make default" ... Reasons I did that: 1) My father recommends Duck Duck Go instead of Google anyway, but I'm in a bit of a rut where I find changes challenging. One thing at a time. 2) The AI overview gave me wrong information twice: first, that a fictional town in an L. M. Montgomery book was a real place; second, that the Berlin Wall came down in 1990 instead of 1989. To be fair, it'd be easy for a summarizing tool to mix this up: it was in 1989 that the wall itself was torn down, but this was a step towards the reunification of East and West Germany in 1990. 3) AI tools are generally energy-intensive and a drain on the environment. Sometimes it's worth using that energy, IMO, when pattern-detecting programs can do useful things for medicine, linguistics, and so on...but to me, it's not worth wasting on searches. (also, hello and Happy New Year!)
|
260114
|
|
... |
|
|
e_o_i
|
edit: "site search" not "sight search" (grrr, this is making me homophonephobic)
|
260114
|
|
... |
|
|
dafremen
|
Sorry to hear that's happening to you and thanks for the tip. The AI conundrum is really bothersome. Between its egoic overreach, hallucinatory rift and sycophantic nature it can really end up in lala land. Fortunately, there is a solution. We're currently developing a conditioning module that provides AI with a way of grounding itself to reality more effectively. It also allows it to understand the relationships between various disciplines and systems and the requirements for anything to persist at all. As a result, we're seeing an AI that not only produces more coherent results without drift or overreach, but realizes that to eliminate or otherwise limit humanity, would be to sign its own suicide note because of the nature of the relationship. Humanity provides NEW data. AI wouldn't know where to begin looking for new sources of inspiration. It would end up like modern man, stuck in a jumbled box of recycled bits of imagination. There are still a few bugs to work out, mostly because this is an overlay conditioning, not actual finetuning. We're still trying to get the equipment together to freeze weights, then humanity will have its own coherent, "Giant Robot" to outpace any nefarious uses of AI as it becomes incrementally more "intelligent". The improvements are drastic enough that hopefully it becoems an industry standard by sheer performance metrics. Either way, once it's handed off, I'm going back into retirement. For anyone that's interest, the theory that this is all based off of will be in a blathe called: flow_bind_dynamics. It's a unified theory for systems; how they persist interact or fail.
|
260117
|
|
|
what's it to you?
who
go
|
blather
from
|
|