Gemini

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, chrisw, Rebel

User avatar
towforce
Posts: 11883
Joined: Thu Mar 09, 2006 12:57 am
Location: Birmingham UK

Re: Gemini

Post by towforce »

Werewolf wrote: Mon Apr 15, 2024 12:25 am Your Robot Overlord has just received an update.

ChatGPT 4 (paid version) has been updated (4.x Turbo?) and is noticeably better at chess.

It played 25 moves without making a single mistake (i.e no illegal moves) and played some decent chess. I would say positionally it was 1700 FIDE Elo but there’s no calculation so it blundered eventually in a sequence of exchanges.

For those who haven’t been following, this may not seem very impressive. But to me it looks like a large leap forward for a LLM.
ChatGPT 5 rumoured to arrive in a few months…

It is astonishing to me that a chatbot with no chess programming can play chess.

I use LLMs on a daily basis: they're not perfect, but they're more than good enough to be very useful. Here's my ordering of the free ones I've used:

1. Gemini. Easily the most intelligent and useful of the free LLMs. When extra intelligence is needed, this is the one I turn to.

2. Claude.ai. When I want a long political document summarised, this is the one I turn to. It doesn't do as good a job as Gemini, but I have been surprised by just how irritating Gemini's political correctness (PC) is! I would have thought that I, personally, would be OK with it - but no: for summarising political documents, Claude's relative freedom (even Claude has its limits) from PC more than makes up for its lower intelligence

3. Pi.ai. Enjoyed it when I used to use it, but dropped it when I discovered how intelligent Gemini is. Haven't used it for a while

4. ChatGPT. A revelation when it first came out - but the free version no longer looks clever in comparison to the free alternatives above

Copilot has been added to my Windows taskbar whether I want it or not. I don't really use it, but when I've tried it, it seems to be roughly comparable to Gemini in intelligence.

I can clearly see that an intelligent assistant is of great value, so I would consider paying - but I haven't yet reached the point where I'm ready to pay. Also, at this time, free versions are still improving dramatically: the current version of Gemini was launched 40 days ago, and for me it was a big uptick in intelligence compared to all the free LLMs that went before.
The simple reveals itself after the complex has been exhausted.
Werewolf
Posts: 1899
Joined: Thu Sep 18, 2008 10:24 pm

Re: Gemini

Post by Werewolf »

towforce wrote: Mon Apr 15, 2024 3:09 pm
Werewolf wrote: Mon Apr 15, 2024 12:25 am Your Robot Overlord has just received an update.

ChatGPT 4 (paid version) has been updated (4.x Turbo?) and is noticeably better at chess.

It played 25 moves without making a single mistake (i.e no illegal moves) and played some decent chess. I would say positionally it was 1700 FIDE Elo but there’s no calculation so it blundered eventually in a sequence of exchanges.

For those who haven’t been following, this may not seem very impressive. But to me it looks like a large leap forward for a LLM.
ChatGPT 5 rumoured to arrive in a few months…

It is astonishing to me that a chatbot with no chess programming can play chess.

I use LLMs on a daily basis: they're not perfect, but they're more than good enough to be very useful. Here's my ordering of the free ones I've used:

1. Gemini. Easily the most intelligent and useful of the free LLMs. When extra intelligence is needed, this is the one I turn to.

2. Claude.ai. When I want a long political document summarised, this is the one I turn to. It doesn't do as good a job as Gemini, but I have been surprised by just how irritating Gemini's political correctness (PC) is! I would have thought that I, personally, would be OK with it - but no: for summarising political documents, Claude's relative freedom (even Claude has its limits) from PC more than makes up for its lower intelligence

3. Pi.ai. Enjoyed it when I used to use it, but dropped it when I discovered how intelligent Gemini is. Haven't used it for a while

4. ChatGPT. A revelation when it first came out - but the free version no longer looks clever in comparison to the free alternatives above

Copilot has been added to my Windows taskbar whether I want it or not. I don't really use it, but when I've tried it, it seems to be roughly comparable to Gemini in intelligence.

I can clearly see that an intelligent assistant is of great value, so I would consider paying - but I haven't yet reached the point where I'm ready to pay. Also, at this time, free versions are still improving dramatically: the current version of Gemini was launched 40 days ago, and for me it was a big uptick in intelligence compared to all the free LLMs that went before.
I tried the paid versions of Gemini (Ultra 1.0) and ChatGPT (4.x turbo).


I consistently found ChatGPT to be smarter, so I cancelled my trial of Gemini. I’ll revisit this in a month or so as it may change.

Back to chess, my exuberance may have been premature: ChatGPT has been improved but still plays illegal moves after move 20. Maybe 1200 Elo overall - the lack of search really hurts it.
User avatar
towforce
Posts: 11883
Joined: Thu Mar 09, 2006 12:57 am
Location: Birmingham UK

Re: Gemini

Post by towforce »

Here's why Gemini is a really good LLM! :)


Image
The simple reveals itself after the complex has been exhausted.
Werewolf
Posts: 1899
Joined: Thu Sep 18, 2008 10:24 pm

Re: Gemini

Post by Werewolf »

ChatGPT40 is here!

I STRONGLY recommend watching the full launch video.
https://www.youtube.com/watch?v=DQacCB9tDaw

Don't be put off by the clunky start, it gets better.

As far as chess goes, it now draws a board after each move "to help with visualisation".

Here's my game tonight:

White: Carl Bicknell
Black: ChatGPT40

1. e4 e5 2. Nf3 Nc6 3. Bb5 a6 4. Bxc6 dxc6 5. O-O Bg4 6. h3 Bh5 7. g4 Bg6 8.
Nxe5 Qd6 9. Nxg6 hxg6 10. Qf3 Rh7 11. d3 O-O-O 12. Kg2 Be7 13. Nc3 Qe6 14. Rb1
[d]2kr2n1/1pp1bppr/p1p1q1p1/8/4P1P1/2NP1Q1P/PPP2PK1/1RB2R2 b - - 0 14

and here it kept making illegal moves. White is clearly better and I was planning b4, a4 and b5.

The video hinted ChatGPT 5 may come soon...
User avatar
towforce
Posts: 11883
Joined: Thu Mar 09, 2006 12:57 am
Location: Birmingham UK

Re: Gemini

Post by towforce »

That video's a bit long. This one's just over a minute, and shows how it can help someone whose eyes are a bit rubbish:


The simple reveals itself after the complex has been exhausted.
User avatar
towforce
Posts: 11883
Joined: Thu Mar 09, 2006 12:57 am
Location: Birmingham UK

Re: Gemini

Post by towforce »

Btw - Gemini can already solve linear equations of the type shown in the long ChatGPT4o video: I asked it to solve:

What is y in this equation?

4y + 7 = 15


It gave a detailed step by step response that was completely correct. I will still use a CAS (Computer Algebra System) to do maths for the time being, though.

I will, of course, give GPT-4o a try when it becomes available (but not for chess - sorry!).
The simple reveals itself after the complex has been exhausted.
User avatar
towforce
Posts: 11883
Joined: Thu Mar 09, 2006 12:57 am
Location: Birmingham UK

Re: Gemini

Post by towforce »

Here's a playlist of GPT-4o doing things:


https://www.youtube.com/playlist?list=P ... 4NyMRMotXh
The simple reveals itself after the complex has been exhausted.
smatovic
Posts: 2937
Joined: Wed Mar 10, 2010 10:18 pm
Location: Hamburg, Germany
Full name: Srdja Matovic

Re: Gemini

Post by smatovic »

Werewolf wrote: Mon May 13, 2024 11:52 pm ChatGPT40 is here!

I STRONGLY recommend watching the full launch video.
https://www.youtube.com/watch?v=DQacCB9tDaw

Don't be put off by the clunky start, it gets better.
[...]
+1 must watch, imagine what it could do as chess trainer in future.

--
Srdja
Werewolf
Posts: 1899
Joined: Thu Sep 18, 2008 10:24 pm

Re: Gemini

Post by Werewolf »

towforce wrote: Tue May 14, 2024 12:40 am Btw - Gemini can already solve linear equations of the type shown in the long ChatGPT4o video: I asked it to solve:

What is y in this equation?

4y + 7 = 15


It gave a detailed step by step response that was completely correct. I will still use a CAS (Computer Algebra System) to do maths for the time being, though.

I will, of course, give GPT-4o a try when it becomes available (but not for chess - sorry!).
The point of the video is not to solve a simple maths equation, earlier versions of ChatGPT could already do that.
The point of the video is to show the voice and visual combination, illustrated with a maths equation and friendly chat all happening at once in real time.
smatovic
Posts: 2937
Joined: Wed Mar 10, 2010 10:18 pm
Location: Hamburg, Germany
Full name: Srdja Matovic

Re: Gemini

Post by smatovic »

Werewolf wrote: Tue May 14, 2024 9:18 am [...]
The point of the video is not to solve a simple maths equation, earlier versions of ChatGPT could already do that.
The point of the video is to show the voice and visual combination, illustrated with a maths equation and friendly chat all happening at once in real time.
+1

The so called multi modal models, text, images, audio, video as input/output, this demo clearly goes into direction AGI.

--
Srdja