Google Releases Its Own 'Reasoning' AI Model (techcrunch.com)
(Thursday December 19, 2024 @10:30PM (BeauHD)
from the rise-of-the-machines dept.)
- Reference: 0175704721
- News link: https://tech.slashdot.org/story/24/12/19/2235220/google-releases-its-own-reasoning-ai-model
- Source link: https://techcrunch.com/2024/12/19/google-releases-its-own-reasoning-ai-model/
An anonymous reader quotes a report from TechCrunch:
> Google has [1]released what it's calling a new "reasoning" AI model -- but it's in the experimental stages, and from our brief testing, there's certainly room for improvement. The new model, called Gemini 2.0 Flash Thinking Experimental (a mouthful, to be sure), is available in AI Studio, Google's AI prototyping platform. A model card describes it as "best for multimodal understanding, reasoning, and coding," with the [2]ability to "reason over the most complex problems " in fields such as programming, math, and physics. [...]
>
> Built on Google's recently announced Gemini 2.0 Flash model, Gemini 2.0 Flash Thinking Experimental appears to be similar in design to [3]OpenAI's o1 and other so-called reasoning models. Unlike most AI, reasoning models effectively fact-check themselves, which helps them avoid some of the pitfalls that normally trip up AI models. As a drawback, reasoning models often take longer -- usually seconds to minutes longer -- to arrive at solutions. Given a prompt, Gemini 2.0 Flash Thinking Experimental pauses before responding, considering a number of related prompts and "explaining" its reasoning along the way. After a while, the model summarizes what it considers to be the most accurate answer.
[1] https://x.com/JeffDean/status/1869789813232341267
[2] https://techcrunch.com/2024/12/19/google-releases-its-own-reasoning-ai-model/
[3] https://tech.slashdot.org/story/24/09/12/1717221/openai-releases-o1-its-first-model-with-reasoning-abilities
> Google has [1]released what it's calling a new "reasoning" AI model -- but it's in the experimental stages, and from our brief testing, there's certainly room for improvement. The new model, called Gemini 2.0 Flash Thinking Experimental (a mouthful, to be sure), is available in AI Studio, Google's AI prototyping platform. A model card describes it as "best for multimodal understanding, reasoning, and coding," with the [2]ability to "reason over the most complex problems " in fields such as programming, math, and physics. [...]
>
> Built on Google's recently announced Gemini 2.0 Flash model, Gemini 2.0 Flash Thinking Experimental appears to be similar in design to [3]OpenAI's o1 and other so-called reasoning models. Unlike most AI, reasoning models effectively fact-check themselves, which helps them avoid some of the pitfalls that normally trip up AI models. As a drawback, reasoning models often take longer -- usually seconds to minutes longer -- to arrive at solutions. Given a prompt, Gemini 2.0 Flash Thinking Experimental pauses before responding, considering a number of related prompts and "explaining" its reasoning along the way. After a while, the model summarizes what it considers to be the most accurate answer.
[1] https://x.com/JeffDean/status/1869789813232341267
[2] https://techcrunch.com/2024/12/19/google-releases-its-own-reasoning-ai-model/
[3] https://tech.slashdot.org/story/24/09/12/1717221/openai-releases-o1-its-first-model-with-reasoning-abilities
Dupe? (Score:2)
by cstacy ( 534252 )
Isn't this story a dupe from like 5 minutes ago?
Or are all the AI stories just so much the same...
Re: (Score:2)
by taustin ( 171655 )
> Or are all the AI stories just so much the same...
Looks to me like the output of the same prompt on each new AI.
"Reasoning" (Score:2)
by VeryFluffyBunny ( 5037285 )
Apparently for GPT LLMs, "reasoning" means "regurgitating the language that people typically use when they thing out loud." Have they actually produced evidence of reasoning, I mean for non-linguistic tasks. That'd be a more realistic task for a "thinking machine."
Better to replace management (Score:2)
If you think about it, these "reasoning" modules should replace the idiot management who keep making these stupid decisions.