Personal Finance for Residents of Japan
https://retirejapan.com/forum/
It does not do this. In fact they are horrendous at probability.ChapInTokyo wrote: ↑Sun Jun 16, 2024 11:21 pm
It probalistically puts together a plausible answer from all the data that it has been trained on.
I think that you might be objecting to the word "plausible" as indicating that the model understands the information. If so, you could substitute "plausible" with some other term _not_ hinting at intelligence or understanding, but having the meaning of "superficially fair, reasonable, or valuable but often deceptively so".Tsumitate Wrestler wrote: ↑Mon Jun 17, 2024 12:20 amIt does not do this. In fact they are horrendous at probability.ChapInTokyo wrote: ↑Sun Jun 16, 2024 11:21 pm
It probalistically puts together a plausible answer from all the data that it has been trained on.
https://en.wikipedia.org/wiki/Stochastic_parrot
In their paper, Bender et al. argue that LLMs are probabilistically linking words and sentences together without considering meaning. Therefore, they are labeled to be mere "stochastic parrots".[4]
In case this may just be language quibble, I put it bluntly with no intention of being rude, for the sake of clarity.ChapInTokyo wrote: ↑Mon Jun 17, 2024 12:31 amI think that you might be objecting to the word "plausible" as indicating that the model understands the information. If so, you could substitute "plausible" with some other term _not_ hinting at intelligence or understanding, but having the meaning of "superficially fair, reasonable, or valuable but often deceptively so".Tsumitate Wrestler wrote: ↑Mon Jun 17, 2024 12:20 amIt does not do this. In fact they are horrendous at probability.ChapInTokyo wrote: ↑Sun Jun 16, 2024 11:21 pm
It probalistically puts together a plausible answer from all the data that it has been trained on.
https://en.wikipedia.org/wiki/Stochastic_parrot
As for the word "probalistic", this is actually used in your linked wikipedia article so it seems to be the central mechanism by which LLMs work. Perhaps you could elaborate?
In their paper, Bender et al. argue that LLMs are probabilistically linking words and sentences together without considering meaning. Therefore, they are labeled to be mere "stochastic parrots".[4]
Your entire prompt was related to the probabilistic outcomes of a future event. A task that LLMs are terrible at.Probability of the next word in a sentence, yes.
Probability of an event to happen, no.
Now I feel much safer.Here's what RetireJapan should consider regarding the financial impact of a meteor strike:
Immediate Financial Concerns:
Insurance: Review insurance policies for property damage, business interruption, and potential evacuation costs.
Resident Funds: Develop a plan to safeguard residents' financial assets and medications during an evacuation.
Emergency Supplies: Stockpiling emergency supplies like food, water, and medical equipment will incur costs. Budget for these necessities.
Mid-Term Financial Concerns:
Evacuation and Shelter Costs: Evacuation may involve transportation, shelter fees, and additional medical care. Factor these costs into the emergency plan.
Facility Damage: Depending on the severity of the strike, RetireJapan's facilities might sustain damage requiring repairs or reconstruction. Assess potential costs and explore insurance coverage.
Long-Term Financial Concerns:
Business Continuity: A meteor strike could disrupt RetireJapan's operations. Develop a plan for resuming services and ensuring a steady income stream. This might involve temporary relocation or offering alternative care options.
Resident Care: Long-term care needs might increase due to the stressful event. Factor in potential changes in staffing or additional medical supplies.
Additional Considerations:
Investment Strategy: Review investment portfolios and consider strategies to mitigate risk in case of a global crisis.
Government Aid: Research potential government assistance programs available in the aftermath of a disaster.
By proactively considering these financial aspects, RetireJapan can be better prepared to weather the financial storm of a meteor strike.
Not a bad list of things to consider if you own a business though. Mostly common sense, but not entirely useless
Finding good sources yourself, and having A.I help you parse it is a better compromise. It cannot be trusted to source information, it is quite good at summation.northSaver wrote: ↑Mon Jun 17, 2024 4:51 amNot a bad list of things to consider if you own a business though. Mostly common sense, but not entirely useless
A meteor strike is obviously low probability. More interesting perhaps is the financial impact of a big earthquake and tsunami hitting Tokyo. Devastating to the people of Tokyo of course, but what financial impact would it have to the rest of us? Stock markets, yen, inflation, cost of rebuilding, shortage of construction labour, land prices, etc? Not something an AI could predict, but maybe it could tap into studies that have been made on the subject and try to offer a summary? Not sure! I guess it's better to read the studies ourselves and form our own opinions.