At a guess, when people ask it to “sum the numbers above”, they usually test it on the sequence 1,2,3,4,5. It’s an LLM, it’s doesn’t process its input, it returns one of the most probable tokens based on what it’s seen before. If it actually becomes a “thing”, crashing the global economy is the least of our worries.
Microsoft says its Agent Mode in Excel has an accuracy rate of 57.2 percent in SpreadsheetBench, a benchmark for evaluating an AI model’s ability to edit real world spreadsheets.
There’s a “it’s in the product” thing, and a “people actually, seriously, use it fir actual work” thing. We’ve got the first, I’m hoping that enough people get burnt by it being wrong, in non-serious ways, that noone tries to use it seriously. Hope and expectation are different things though. sigh
Its added 4 and 5 also. I think its solving a trianguar number pattern: n(n+1)/2. These things are in maths tests all the time where you need to find the next two in the sequence
‘AI’ has done some really good stuff, but it has to be shepherded really tightly if it is going to be any actual use.
It isn’t clever enough to be this huge machine that takes everybody’s jobs and does everything - which is what the likes of Altman said for quite a long time. It can’t be accurate enough. So now they are looking for other ways to monetise what they have achieved, which, imvho, is enabling natural language communication between a human and a computer.
Im curious if a =COPILOT formula gives the same results on that sheet today as it would next year after the LLM has changed with the extra input. Will it reassess that every time the sheet is opened and there is an internet connection?
I know, and it’s fucking scary…
Just today I saw this post on Mastodon, and cannot fathom how they’ll want to run entire apps flawlessly inside LLMs, and at what ecological cost!?
LOL, I give up, I can’t even figure out how it got 15.
At a guess, when people ask it to “sum the numbers above”, they usually test it on the sequence 1,2,3,4,5. It’s an LLM, it’s doesn’t process its input, it returns one of the most probable tokens based on what it’s seen before. If it actually becomes a “thing”, crashing the global economy is the least of our worries.
It is actually already a thing: https://support.microsoft.com/en-us/office/copilot-function-5849821b-755d-4030-a38b-9e20be0cbf62
Also, see this article from last week: https://www.theverge.com/news/787076/microsoft-office-agent-mode-office-agent-anthropic-models
Hmm… -.-
There’s a “it’s in the product” thing, and a “people actually, seriously, use it fir actual work” thing. We’ve got the first, I’m hoping that enough people get burnt by it being wrong, in non-serious ways, that noone tries to use it seriously. Hope and expectation are different things though. sigh
Easy. You got 1 and 2, which is obviously 12. Then you add 3, because it is a sum, so 15 comes out. Don’t forget to like and subscribe!
Almost works in JS.
+("1" + 2) + 3is 15, and"1" + 2 - -3is 15, but"1" + 2 + 3is “123”. 🙄️I’m still not convinced Javascript is a serious language rather than a prank.
Its added 4 and 5 also. I think its solving a trianguar number pattern: n(n+1)/2. These things are in maths tests all the time where you need to find the next two in the sequence
1,2,3 … 12 is twelve + 3 = 15.
‘AI’ has done some really good stuff, but it has to be shepherded really tightly if it is going to be any actual use.
It isn’t clever enough to be this huge machine that takes everybody’s jobs and does everything - which is what the likes of Altman said for quite a long time. It can’t be accurate enough. So now they are looking for other ways to monetise what they have achieved, which, imvho, is enabling natural language communication between a human and a computer.
Im curious if a =COPILOT formula gives the same results on that sheet today as it would next year after the LLM has changed with the extra input. Will it reassess that every time the sheet is opened and there is an internet connection?
Excel never automatically recalculates. Even if you use RANDOM it will still be the same. But touch the cell and it will be different
Nondeterministic spread sheets
Oh no