Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Bard is now Gemini, and we’re rolling out a mobile app and Gemini Advanced

blog.google

Post reply on HN
Print view View original

141–150 of 1001 posts

Re: Bard is now Gemini, and we’re rolling out a mobile app and Gemini Advanced

#141

Post by OJFord »

shantara wrote: post #100
alphabetting wrote: post #83

Gemini Ultra seems better on logic than GPT4. Still messing around testing but here's a prompt Ultra nailed but GPT4 completely botched: Tabitha likes cookies but not cake. She likes mutton but not lamb, and she likes okra but not squash. Following the same rule, will she like cherries or pears https://i.imgur.com/KW6gQbc.jpeg https://i.imgur.com/OSHSvLp.png

I would have never guessed the answer. With such little data available, one can invent any arbitrary rules to fit their favorite answer. It would be more impressive to practical use cases, if a LLM simply said that it's impossible to guess without inventing their own reasoning or looking up the answer online.

Same, I had to look to see what the intended answer was.

In fairness though, GPT4 was objectively incorrect, it's not even internally consistent or coherent - it either thinks b & h are vowels, or that lamb and squash don't end in those letters, or has changed its mind about the rule mid-sentence, or something.

Re: Bard is now Gemini, and we’re rolling out a mobile app and Gemini Advanced

#142

Post by nycdatasci »

alphabetting wrote: post #117
rvnx wrote: post #97

Earlier quoted context omitted.

Proof of Gemini cheating: https://i.imgur.com/eYJDFjS.png Answer about cherries falling from the sky... (there is no question or context beforehand, this is the first question of the chat)

That's a bummer. I just made this one up which GPT-4 failed and Gemini got right but definitely need to do more testing to see what's being taken from online and what is actual reasoning. https://i.imgur.com/3sNr3LW.png https://i.imgur.com/EIj0nZg.png Edit: When I did your prompt it got it right on the first try fwiwi https://i.imgur.com/E3zYEca.png

Great example! Thanks for sharing. I tested the same prompt using a variety of GPT-4 model versions and it consistently got the answer wrong. When I continued with "use algebra/formulas to confirm your answer" it corrected the initial mistake, but I'm surprised that it's not doing that out of the box.

Re: Bard is now Gemini, and we’re rolling out a mobile app and Gemini Advanced

#143

Post by asystole »

nolok wrote: post #116
nazgul17 wrote: post #77

Earlier quoted context omitted.

No, you would assign a probability to the event. And that probability would be higher for a Google product.

I would disagree with you, my personnal anecdotal data showing google to be more reliable at it than others. In the lack of hard properly collected data, this give all of them the same probability. I'm sure someone out there thought "let's use microsoft skydrive because google is unreliable".

[flagged]

Re: Bard is now Gemini, and we’re rolling out a mobile app and Gemini Advanced

#144

Post by iandanforth »

I reran a few of the queries that Bard had failed on in Gemini/Ultra and didn't see any improvement. Made the same, or new, logical errors, hallucinated facts, failed to recognize things I was describing etc. It did do better on recognizing an image I uploaded but it went from accurate to nonsensical in the same response.

Re: Bard is now Gemini, and we’re rolling out a mobile app and Gemini Advanced

#145

Post by druskacik »

Prompt: Spell the word lollipop backwards.

Gemini: The word "lollipop" spelled backwards is: popillol I hope this sweet treat of a word brightens your day!

I'm impressed. However, it still fails on "How many words are in your response to this?".

Re: Bard is now Gemini, and we’re rolling out a mobile app and Gemini Advanced

#146

Post by jjackson5324 »

dobladov wrote: post #121
dobladov wrote: post #119

I would be willing to pay for an AI that it's not constantly worrying about my safety and ethics, it's very frustrating that even when I ask for tech related topics, Gemini it's holding information on the basis of actions that might cause harm to my devices.

As expected, I asked the same questions to perplexity.ai, and it gave a correct response, just with a minor disclaimer at the end explaining the risks.

It's not even funny how much better perplexity.ai is than everyone else

Re: Bard is now Gemini, and we’re rolling out a mobile app and Gemini Advanced

#147

Post by antonioevans »

I'm not impressed. When comparing Gemini/Bard to ChatGPT + GPTs, Bard/Gemini feel more like a search engine. I asked Gemini for help in planning a date with my date, who is famous enough that GPT4 knows her and her art. However, Gemini immediately started giving me step-by-step instructions to plan the date. I had to tell it to slow down and ask me questions first before giving an answer. It complied this time, but after the Q&A, it provided nearly the same response as before, without any personalization. Next, I asked about my artist friend, but Gemini had no clue. I even said, "come on, you have to know her," but it simply repeated that it didn't know her. Another issue I encountered was with images. I tried sending a few, but Gemini couldn't describe them. I spend around 8-10 hours a day playing with LLMs, but so far, I'm not impressed.

Re: Bard is now Gemini, and we’re rolling out a mobile app and Gemini Advanced

#148

Post by IanCal »

menschmanfred wrote: post #62

Mh weird that they bundle it with storage. Makes downgrading hard if you don't track your usage and get used to it

Even weirder is that it's only bundled with 2TB of storage. If you have more than that you cannot upgrade without losing storage space.

Re: Bard is now Gemini, and we’re rolling out a mobile app and Gemini Advanced

#149

Post by flerchin »

c-fe wrote: post #102
ipsum2 wrote: post #93

Earlier quoted context omitted.

Note that Gemini pulled the answer off the Internet, while GPT-4 didn't. The answer can easily be found via Google search. Changing up the question a little, I reversed it and asked Ultra and it was unable to answer: Jake likes coke but not pepsi. He likes corn but not popcorn, and he likes pens but not pencils. Will Jake like salmon or cheese? https://i.imgur.com/lWU9HHS.png edit: why was this downvoted? I don't und…

I dont think your reversed question makes sense. In the OP example, one item was always smaller/younger then the other item. In your example, I cannot, even as a human, identify the differences

> even as a human

Would all the humans please take one step forward. Not so fast c-fe.

Re: Bard is now Gemini, and we’re rolling out a mobile app and Gemini Advanced

#150

Post by firstrowraver »

I tested it immediately, but it is disappointing. At least here in Switzerland, it is not able to generate images, and a simple "look up this website and summarise the content" does not work either (can't access the website, but its a public website without any crawling limitations). I don't understand why Google is launching a product like this.
Post reply on HN