Staff can't code? No prob. Singapore superapp's LLM whips up apps for them

Southeast Asia's Uber-esque superapp, Grab, has developed a tool that allows its employees to build large language model (LLM) apps without coding. The company this week described the tool, which it calls Spellvault, as empowering "everyone in the company, even those without coding skills," to "create LLM apps that have access …

[Bebu sa Ware]

Like fungi on the forest floor...

AI/LLM are like toadstools and mushrooms popping up all over the place. Apart from being potentially lethal to consume they have already sent out their far reaching tentacles (mycelia) to feed in decaying wood.

A Nick Leeson digital twin in the offing?

[sarusa]

Well sure?

I can easily crap out new LLM apps in minutes too by forking existing stuff and changing a line of code. How well do their new automatically shat out apps actually work?

... and well, maybe it doesn't even matter? Most of the people deploying them don't seem to care at all. Most LLMs just fountain piss randomly and punters happily guzzle it and the C-levels roll around in and lick up what comes out of them.

So I guess most of the 'AIs' their 'AI' is producing are the usual 'AI' crap, but that crap sells? Hooray?

[ChoHag]

> The AI framework integrates a retrieval step before generating responses, which is claimed to improve accuracy and relevance of LLMs.

So it does your web search then wraps the results up in bullshit before presenting them to you? How many trillions?

[T. F. M. Reader]

Doesn't compute?

Users "with no coding skills" "can specify APIs if the model needs additional data from external sources".

See Title.

[An_Old_Dog]

Tripping Up LLM Apps / GIGO

The part where the user uploads relevant data to the DataVault is the sticking point. If the user is unaware of all the relevant data and/or chooses not to upload it all, the resultant app will be crippled, and will output garbage.