reply on: NotebookLM Workshop \ stacker news ~AI

pull down to refresh

54 sats \ 2 replies \ @optimism 25 Jun \ on: NotebookLM Workshop AI

Thanks for this awesome report.

Brainstorming sessions - if you have a good group - can be really awesome and trigger you into whole new ways of thinking (just like interacting on low-slop online places like SN or Github can if you can keep an open mind and are willing to pick your battles.)

I noticed a thing on the padlet. It says for your first interaction: "Think about one topic that you would like to learn about". However, that is not how one judges the usability of cognitive solutions. Instead, think about the topic you are most confident about your own knowledge on.

Here's how I tested NotebookLLM back in the day: I fed it a pdf that I authored myself on a controversial topic, and then watched what it made of it and queried it. It was very scary scary to see that it seemingly blended outside popular beliefs into my super awesome non-biased paper that didn't mention some of the statements the "podcast" made and how some of my assertions were transformed into a slippery slope of what people have written on the interwebs. This is why I personally have avoided it since.

100 sats \ 1 reply \ @cryotosensei OP 20h

I don’t know if me saying this will make you shake your head in disdain, but all public teachers in Singapore can use their official email address to register an account with NotebookLM. I assume doing so yields me more privileges than registering with my throwaway personal email. As such, the thought of verifying the credibility/inherent biases never crossed my mind. Neither will it cross my colleagues’ minds I reckon haha

What was that controversial topic that forced NotebookLM to tip its hand?

50 sats \ 0 replies \ @optimism 20h

the thought of verifying the credibility/inherent biases never crossed my mind

I think most people do not verify, but "trust". When you're working as a developer on Bitcoin systems, or of any critical codebase really, red teaming something to oblivion is the default mode. Think SavageLinus(tm) shouting on the kernel mailing list: "WE DO NOT BREAK USERSPACE!" Which in normie talk means: don't eff up.

Now, if you shouldn't eff up, it means you must view every change from the perspective of the enemy (this is what red teaming is) and if you're changing your cognitive tooling by introducing something like NotebookLLM, this is super important: the tool is there to help improve your output. Worse quality cognition is not a good outcome.

What was that controversial topic that forced NotebookLM to tip its hand?

In this case it was about a particular Bitcoin soft fork I wrote a piece on in the past. So it had information (and emotional context!) that was more recent than my piece and it came back in both the "podcast" and when querying. So this felt like contamination from the training data to me and thus I wrote it off.