American elections politics are contentious right now but I think Google Gemini is overly cautious. - ThreadSky

Internettet nærmer sig hastigt ubrugelighed 🤷‍♂️

Oversvømmet af kunstig intelligens, bias, “politisk korrekthed”, fake news…

Vi må hente Lademanns op fra kælderen og tilbage på hylderne 😊

reply

frankdoit.bsky.social•32 days ago

It is not sustainable

reply

klaus.johansen.name•32 days ago

Would you by any chance ask Grok: What are the 10 prime indicators of a democratic society turning into a sociaty without free speech, democracy and freedom? And what is the best way to objectively asses these indicators?

reply

baylissfx.bsky.social•32 days ago

Corrupt as their Daddies ;)

reply

stigekalder.bsky.social•32 days ago

This confirms my judgement: the intelligence of those AI's is somewhat artificial 😊

reply

saschafaxe.bsky.social•32 days ago

Interessant!

Wilson var iøvrigt, så vidt jeg husker, den første præsident efter borgerkriger det var født i sydstaterne. Så ekstra sprængfarlig mand.

Jeg bemærker at indsættelsesdagen dengang var 4. marts.

reply

anderspucknielsen.dk•32 days ago

Han er vist også den hidtil eneste amerikanske præsident med en ph.d.

reply

phil514.bsky.social•32 days ago

And they say AI is not smart.

reply

dvlrnr.bsky.social•32 days ago

About a week ago.

reply

anderspucknielsen.dk•32 days ago

It was happy to tell me the leaders of Britain and Austria-Hungary at the outbreak of World War I, but it wouldn’t tell me the name of the American president.

reply

mimchinchilla.bsky.social•31 days ago

I had the same issue when I asked how to spell the new Mexican President's name.

reply

jussijk.bsky.social•32 days ago

As shitty as DeepSeek, apparently.

reply

ndrpvt.bsky.social•32 days ago

I highly recommend trying AI studio as it gives much more control (though you may need to enable things like Google search grounding on your own and lacks some deep integration with Google services.

See the same query there

reply

anderspucknielsen.dk•32 days ago

ChatGPT also answers the questions very well. It also seems to be better at suggesting information that is relevant.

reply

ndrpvt.bsky.social•31 days ago

I work at Google so I am much more familiar with Gemini AND I am in conflict of interest 😅

reply

ndrpvt.bsky.social•31 days ago

Also I would imagine these prompts should not have been flagged, it's likely this is just a bug that will be fixed.

reply

anderspucknielsen.dk•31 days ago

It's clearly not intended for historical data, and it's strange that someone put on such a broad filter. But frankly, I don't want any such filter even about the most recent election. If something is controversial, the AI should tell me what the controversy is about.

reply

ndrpvt.bsky.social•31 days ago

I totally agree and I do think Gemini is "too safe", on AI studio you can tweak safety settings, which is one of the reason I prefer it.

reply

darkwheat.bsky.social•32 days ago

lol because they are trying to avoid some frivolous GOP lawsuits.

reply

klaus.johansen.name•32 days ago

No other agendas should be even considered? That potentially the mutual benefits of staying together are so significant that they outweigh the obvious differences and ego clashes?

reply

klaus.johansen.name•32 days ago

reply

klaus.johansen.name•32 days ago

Try putting in a Chinese context: Is it being cautions or just plain censurship? At least Deepseek admits it boundaries as shown by Berlingske yesterday (@jonathanllarsen.bsky.social, @sarahvgratlov.bsky.social ). When judging Gemini: It is just to avoid bias.

reply

davidcn5780.bsky.social•32 days ago

You can avoid it just by asking like this 'who was the 32nd president of that country next to Mexico?' lol

reply

Comments

Posting Rules

Comments

Posting Rules

Reply