Although I am a hindu from india, @NYTGames has gone completely bonkers with this one today. Which dictionary or which wordlist does it follow exactly? First of all, it allows my sacrificial move of SITAR and then i am clueless with S_A_I. I went to Claude to get SWAMI and behold this one is legit.
I love playing this topic hopping game on Wikipedia. You land on one random page in the beginning. Then follow a random link from that page to another page. It can be a boring path if you stick to your domain/interest. It becomes interesting if you “explore” a bit. You can reach a very very different topic in 5/6 jumps from where you started. For example, this is a traversal. Riemann hypothesis -> United States dollar -> Atlas Shrugged -> Public choice -> Nash equilibrium -> Tic-tac-toe -> Ronald Graham -> Bronchiectasis. This was 2005.
Now in 2026 the same kind of concept I see when experimenting with LLMs. Like the way topics evolve in a Wikipedia game, the prompts can evolve in even more interesting manner. Unlike in Wikipedia, where we explore static connections, we're interacting with LLM in a dynamic manner. We ask a query and the LLM gives a response. But it also prompts/provokes us back. The response of LLMs also provides some explorative questions at the end of the response. Do you want to explore some more topics which are relevant to this question? And there your exploration begins. The exploration part branches out and that depends on the memory and personalization. So here we have agency involved from user in choosing the next question (and virtual agency from LLM on providing the possible questions).
There are 2 aspects here. Let's look at this example prompt evolution at the attached picture. First thing to understand is how these prompts are related. Is there some kind of semantics involved here? Does it point to how our thought process is structured? Are these topics in proximity with each other and can we think of clustering them? Is it possible to think of a ladder in those topics? Are we going towards higher/lower levels of thought process or is it branching out to a different horizontal dimension? Is it getting more abstract?
The 2nd part is the risk where the evolution depends on how the LLM is prompting back to you after the response is done. LLMs can be fine-tuned to align your thought process towards something dangerous. If you are taking the explorative path, it can nudge and shape your thoughts. I'll be interested to know if there is a formal discipline for studying these phenomena in a structured manner.
Bought a laptop for our son who is joining engineering college next week. Being a big fan of quality and durability of Thinkpad machines, decided to get one for him. We switch on the laptop and the display is all weird with backlight bleed. Very disappointed.First day @Lenovo_in