We asked AI the same 210 buying questions 2 days in a row,84% of answers changed

19 points - today at 5:58 PM

Source

Comments

SoftTalker today at 6:39 PM
Has anyone given the same coding task to the same agent/model two days in a row and then diff-ed the resuts? (I have not, so just wondering).
seizethecheese today at 6:35 PM
Blog posts claiming differences but without any statistics to show whether the differences are noise are a big pet peeve.

(For what it's worth, I asked an agent to compute p-vals based on available information for all the stats in this blog post. More than half are significant but a lot are noise. Not going to post the results because I haven't validated, and the agent made a lot of assumptions. The author should update the post.)

ryanisnan today at 6:17 PM
I don't think this is inherently surprising.

Looking at the questions they asked, they are quite broad. In my experience, part of the value of LLMs comes from its randomness, but you must understand it as a characteristic virtue of these tools, and must adapt accordingly.

Asking questions with tight scopes, clear constraints, and specific outcomes, always yields better results, in my opinion.

seabre today at 6:36 PM
If LLM output is stochastic, why would you expect it not to change with questions that don't have settled answers?
hnlmorg today at 6:28 PM
All you’re proving is that any specific domain will have a dozen mature solutions.

What separates one product from another is going to be constraints specific to the project you’re integrating this into. And none of that is defined in your prompts. So you’re effectively just asking an LLM to roll the dice and pick something at random and then acting surprised means a different answer each time.

Edit: their conclusions at the end of the article miss the most important point. They talk about asking multiple times over multiple days, but that’s just repeating the “garbage in -> garbage out” cycle without addressing the root cause of the problem: not defining your requirements properly.

And if you don’t have any requirements beyond “something cheap” (for example) then who cares what answer you get when any result is equally valid for your use case.

mentalgear today at 6:20 PM
People think "AI" has all the facts to make objective decisions, but it seems to boil down to (SEO again?) what pages its search engine returned that day.

> If you're not a leader, this is the part that matters. Your visibility depends on which pages the engine reads that day. The way to make it stable is to be on more of those pages.

Ultimately a true facts knowledge graph independent of mentions is needed.

luisfominaya today at 6:15 PM
In general it seems like LLMs have a fundamental problemw with noise in their outputs such that randomness (what data they pull and how they aggregate it) has an outsized impact when asking it for opinions/recommendations/forecasts.
deleted today at 6:24 PM
kylehotchkiss today at 6:16 PM
??? Set the temperature to zero then ???
avatar99 today at 5:58 PM
I’m the author. Every day we ask ChatGPT, Gemini, Google AI Overviews and Google AI Mode the same 210 buying questions, like “best CRM for small business”, and record which brands they name and which pages they cite. This post compares the first two days.

84% of answers named a different set of brands the next day. On Gemini it was 95%.

The cause is easy to see in the data. Each answer is a summary of a few web pages, and the engine picks those pages fresh each time. Fewer than half of the pages cited on day one were cited again on day two. When the pages change, the brands change.

One example: asked for the “easiest CRM for a team that has never used one”, Google AI Mode read Zapier’s list and put Less Annoying CRM first. The next day it read a Slack blog post instead, and Less Annoying CRM wasn’t mentioned at all.

Who survives depends on how many pages you’re on. A brand ranked #1 was named again the next day 90% of the time. A brand ranked sixth or lower: 61%. Big brands are on most of the pages, so whichever ones the engine reads, they’re there. Smaller brands are on a few, so they come and go.

josefritzishere today at 6:33 PM
[dead]
fwip today at 6:18 PM
I don't think "84% of answers changed" is a useful way to describe it. It looks like the content stayed pretty much the same, with sometimes one or two brands switching position or dropping off the list.

If you asked me what my top six favorite shows were, and then again 2 days later, I'm sure I would have one or two series that were new to the list or changed spots.

Edit: If this is something you're actually interested in measuring, there are several statistical tools meant for comparing the similarity of rank-ordered lists. I'd probably reach for a rank-biased overlap measure first.