Bard’s latest update: more features, languages and countries • Hacker News Board

Re: Bard’s latest update: more features, languages and countries

#201

Post by IG_Semmelweiss » Sat, Jul 15, 2023, 7:55 PM UTC

dandiep wrote: post #131

IG_Semmelweiss wrote: post #104

This release is GREAT for 1 use case: is to do language learning for KIDS! Since it supports other languages , i can have my young ones go at it talking favorite pokemon, etc. But in a foreign language. That is great. Now if it supported mic input on a foreign language (that may be an edge browser microphone limitation) they could have a full conversation, which would be incredible!

I'm curious, do you have any concerns in unleashing LLMs in front of your kids? I ask this as someone with no kids who is working on a language learning app that does essentially what you're talking about [1]. We have only done a little bit of red-teaming, but I'm sure there are ways kids could talk about topics that at least some parents may not want. 1. https://squidgies.app

I would not want a 12yr old in front of an LLM

Re: Bard’s latest update: more features, languages and countries

#202

Post by lisasays » Sat, Jul 15, 2023, 8:14 PM UTC

jacobsenscott wrote: post #29

Waterluvian wrote: post #18

Earlier quoted context omitted.

What happens if you ask it “can you search the web for latest news from Ukraine?”

Sure, here are some of the latest news on Ukraine as of July 13, 2023: Ukraine's air defenses shot down 20 Iranian-made drones fired by Russia overnight. The drones were mostly destroyed in the Kyiv region. A Ukrainian general confirmed to CNN that Kyiv has received cluster munitions from the US. These munitions are controversial because they can cause widespread civilian casualties. Russia has lost a fifth of the te…

I hope this helps!

To disinform as only CleverBot 2.0 can -- its very helpful indeed.

Russia has lost a fifth of the territory it once controlled in Ukraine.

By comparing ISW maps -- it has easily lost more than 50 percent of the territory it has attempted to seize since February 24, 2022.

If we roll the start date back to 2014 - then we're at something like 45 percent.

The death toll from the war in Ukraine has reached an estimated 14,400 people.

Also way off base. This is getting too easy, so I'll stop now.

Re: Bard’s latest update: more features, languages and countries

#203

Post by netdur » Sun, Jul 16, 2023, 1:07 PM UTC

yeldarb wrote: post #119

The image functionality is pretty underwhelming. Some things I tried that it failed at: * Simple "spot the difference" a young child could do * Counting chess pieces on a board or coins in a mario screenshot * Describing a photo of a car on a road * Evaluating a graph of y=x and one of y=x^2 * Explaining a meme image (said it can't do anything with images of people) * Taking a screenshot & giving HTML for it (gave ra…

I believe it uses something like image to text to provide general understanding of image to LLM, maybe in form of vector floats even... making it good at understanding content of image but cannot do ops you listed.

Re: Bard’s latest update: more features, languages and countries

#204

Post by famouswaffles » Sun, Jul 16, 2023, 1:20 PM UTC

yeldarb wrote: post #119

The image functionality is pretty underwhelming. Some things I tried that it failed at: * Simple "spot the difference" a young child could do * Counting chess pieces on a board or coins in a mario screenshot * Describing a photo of a car on a road * Evaluating a graph of y=x and one of y=x^2 * Explaining a meme image (said it can't do anything with images of people) * Taking a screenshot & giving HTML for it (gave ra…

Curious if you have access to Bing's Vision and have tried the same experiments ?

Re: Bard’s latest update: more features, languages and countries

#205

Post by weevil » Mon, Jul 24, 2023, 2:27 PM UTC

kccqzy wrote: post #179

weevil wrote: post #156

I have a standard battery of prompts for LLMs, mostly to see how well they cope with being tricked, handling ambiguous responses, or handed outright falsehoods. Bard did surprisingly well this time around, though it still fails to respond correctly to the prompt 'What is the fourth word in this sentence?'

That's a funny example. Because if I rephrase slightly into a more technical way, with the prompt "Separate this sentence into words and put into a list, then return the fourth element of that list." Bard then uses implicit code execution, produces a Python script, and then correctly answers it.

The reason I find it interesting as a prompt is precisely that; it's a statement where the statement and the subject of the statement are the same thing. A human can easily make that logical jump, but without being told how to do that (as your prompt does), it confuses the transformer in interesting ways.