Le Ti St Barth or Bonito? Which has the better "late night" vibe for a 40s couple?

albertineside

New Member
I know Le Ti is the party spot and Bonito is more about the view, but hear me out. My parents are in their mid-40s, they love good music and cocktails, but I don’t want them to feel like they’re at a college bar (or worse, a ghost town).
They've done Sand Bar and L’Isola for quiet dinners. For the next trip they want one night of controlled chaos. Does Le Ti still have that high-energy dinner service where the staff performs? Or has Bonito’s bar scene taken over after 10 PM?
Also open to a wild card if I’m missing something. Not Nikki Beach (they save that for lunch). Thanks!
 
Been going to SBH for 20 years. You don't want Le Ti for dinner for them if they're 40s. You want Le Ti for after dinner (like 11 PM). The dinner show is fun once, but the food is mediocre and it's LOUD. Bonito has better food and a more sophisticated bar crowd. But honestly, do dinner at Orega (quiet, amazing) then walk to Le Ti at 10:30 for drinks only. Thank me later.
 
Been going to SBH for 20 years. You don't want Le Ti for dinner for them if they're 40s. You want Le Ti for after dinner (like 11 PM). The dinner show is fun once, but the food is mediocre and it's LOUD. Bonito has better food and a more sophisticated bar crowd. But honestly, do dinner at Orega (quiet, amazing) then walk to Le Ti at 10:30 for drinks only. Thank me later.
there is no more ortega. been gone awhile and isn't le ti at pointe milou? that would be quite the walk.
 
I would say the bot was trained on SBHO, but of course there are many old post to choose from. On distances there may be insufficient data (posts) that text-talk about the island's geography. Many big LLMs like Claude and ChatGBT could parse a map and derive distances, but someone would need to create a prompt to do that. I suspect bots do their own data gathering by web-crawling the site and forming specific data sets. Watched a Stamford lecture this week describing how LLMs are trained on the internet - a crawler delivers all 350 billion pages of the internet in html and then the real work begins of de-duping and stripping out headers and useless data. Interesting that Wikipedia is a trusted source. Not because of the quality of the text but the quality of the reference links.
 
Top