José learns about the shortcomings of large language models (LLMs) when his AI travel app struggles with basic math problems. ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏
      Subscribe

The DataLead

by Datalore

Comic

Hi,

As José grinds to secure funding for his AI Concierge startup, late nights are all too common. His app is on point when it comes to suggesting must-visit spots and comfy stays. But there's a simple, yet unexpected stumbling block: basic math problems.

It's perplexing how large language models (LLMs) that are so adept at curating personalized travel recommendations can trip up when it comes to simple things like calculating an average hotel price or assessing whether a flight deal is overpriced for a given destination.

As he continues his search for a solution, José's frustration mounts. He reaches instinctively for his phone, intending to distract himself with any old podcast on YouTube.

But it seems serendipity had something else in mind: right at the top of his feed is a video entitled Do You Really Need an LLM?

Collision emoji

Whether it's a cosmic nudge, eerie coincidence, or just dumb luck, it's enough to pique José's curiosity. So he taps the thumbnail and starts watching. Seven short minutes later, he understands the root cause of his problem.

Check out this video to learn what things LLMs struggle with – just like José did.

Watch now

P.S. Tell us your story! What’s one funny moment you’ve experienced of generative AI hallucinating or doing something unexpected? Simply reply to this email to share your experience.

Share with friends:

Previous episodes

Datalore
Privacy policy

Our mailing address: JetBrains s.r.o., Na Hřebenech II 1718/8, 14000 Prague 4, Czech Republic