Science

(found in @keccers.eth replies)

MIT just published the first brain scan study of ChatGPT users… yikes. 

granted it is a 4 month study with a sample size 54 patients (small), but glad someone is studying this.

ChatGPT users had a ~55% reduction in brain connectivity. 

made them 60% faster at completing tasks, but reduced the "germane cognitive load" needed for actual learning by 32%.

also really interesting to see how “search engine” users compare to pure brain vs ChatGPT.

skip to discussion and conclusion here:

The Farcaster formerly known as Kchamp. Books, beats, bodybuilding, bright ideas and beautiful things.

katherinechampagne.com

(found in @keccers.eth replies)

MIT just published the first brain scan study of ChatGPT users… yikes. 

granted it is a 4 month study with a sample size 54 patients (small), but glad someone is studying this.

ChatGPT users had a ~55% reduction in brain connectivity. 

made them 60% faster at completing tasks, but reduced the "germane cognitive load" needed for actual learning by 32%.

also really interesting to see how “search engine” users compare to pure brain vs ChatGPT.

skip to discussion and conclusion here: https://arxiv.org/pdf/2506.08872v1

building farcaster | surfing @ venice, ca | eating @ gjusta / gjelina | “lucky me, lucky mud” 🤠

Makes me wonder if it's just like exercise. Our brain, just like a muscle, needs to be stretched and have applied strain of working through a problem, idea or thought to grow. When using ChatGPT, we check out while the LLM is doing more heavy lifting. 

Curious to know what the studies are when it comes to typing vs. writing. Since writing grooves the brain deeper with the subject matter since it's more kinesthetic 🤔

CEO of RYFT - On a mission to create the future of human possibility through digital and spiritual technology. Visionary @ $FCAST | LMT

Moni Latam. Connecting builders, projects & users. Casts in English & Español.
Channels: /comunidadesweb3 /onchainloteria
Project: CryptoConexion.com

I mean, the sample size is incredibly small, I wouldn't let create a conclusion for an extreme conclusion such as; we need to rely on LLMs less. But rather, we need to continue to value learning in daily lives.

The study doesn't account for qualitative variables like that, does the subject value learning, as well as the human limitation of 10-bits per second. We have a very narrow neurological input threshold and it makes sense that LLM technology eases the bandwidth of that.

Completely agree. Just like with any tool. It's about our approach, intent and awareness.

I think this study is important for us to take a step back and reflect on how much we are relying on these systems. What the efefcts could be. For me, I try to use them to enhance creations and not be the genesis of that creation.

but i think it also how you use chatgpt or llms in general no? when you have LLMs do all the work, yes. I do agree it is like exercise, and this is a reminder on why we have to not rely on LLMs as much as we do (and I do).