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

Gemini 2.5 Deep Think

blog.google

Post reply on HN
Print view View original

71–80 of 259 posts

Re: Gemini 2.5 Deep Think

#71

Post by mettamage »

littlestymaar wrote: post #37
highfrequency wrote: post #3

Approach is analogous to Grok 4 Heavy: use multiple "reasoning" agents in parallel and then compare answers before coming back with a single response, taking ~30 minutes. Great results, though it would be more fair for the benchmark comparisons to be against Grok 4 Heavy rather than Grok 4 (the fast, single-agent model).

That this kind of approach works is good news for local LLM enthusiasts, as it makes Cloud LLM using this more expensive while local LLM can do so for free up to a point (because LLM inference is limited by memory bandwidth not compute, you can run multiple queries in parallel on your graphic card at the same speed as the single one. Until you become compute-bound of course).

Wait, how does this work? If you load in one LLM of 40 GB, then to load in four more LLMs of 40 GB still takes up an extra 160 GB of memory right?

Re: Gemini 2.5 Deep Think

#72

Post by jstummbillig »

asaddhamani wrote: post #19

I find it interesting, how OpenAI came out with a $200 plan, Anthropic did $100 and $200, then Gemini ups it to $250, and now Grok is at $300. OpenAI is the only one that says "practically unlimited" and I have never hit any limit on my ChatGPT Pro plan. I hit limits on Claude Max (both plans) several times. Why are these companies not upfront about what the limits are?

Because if you are transparent about the limits, more people will start to game the limits, which leads to lower limits for everyone – which is a worse outcome for almost everyone.

tldr: We can't have nice things, because we are assholes.

Re: Gemini 2.5 Deep Think

#73

Post by NitpickLawyer »

lynx97 wrote: post #7
highfrequency wrote: post #3

Approach is analogous to Grok 4 Heavy: use multiple "reasoning" agents in parallel and then compare answers before coming back with a single response, taking ~30 minutes. Great results, though it would be more fair for the benchmark comparisons to be against Grok 4 Heavy rather than Grok 4 (the fast, single-agent model).

I am surprised such a simple approach has taken so long to be actually used. My first image description cli attempt did basically that: Use n to get several answers and another pass to summarize.

I have a similar thing built around a year ago w/ autogen. The difference now is models can really be steered towards "part" of the overall goal, and they actually follow that.

Before this, even the best "math" models were RLd to death to only solve problems. If you wanted it to explore "method_a" of solving a problem you'd be SoL. The model would start like "ok, the user wants me to explore method_a, so here's the solution: blablabla doing whatever it wanted, unrelated to method_a.

Similar things for gathering multiple sources. Only recently can models actually pick the best thing out of many instances, and work effectively at large context lengths. The previous tries with 1M context lengths were at best gimmicks, IMO. Gemini 2.5 seems the first model that can actually do useful stuff after 100-200k tokens.

Re: Gemini 2.5 Deep Think

#74

Post by suddenlybananas »

andsoitis wrote: post #56
crowcroft wrote: post #41

Earlier quoted context omitted.

Mainframes are the only viable way to build computers. Micro processors will never figure out how to get small and fast enough for personal computers to reach escape velocity.

Why do you think the analogy hold?

Well it's an analogy, that's a water-tight argument.

Re: Gemini 2.5 Deep Think

#75

Post by panarky »

cpursley wrote: post #20

I can’t even convince Gemini CLI while planning things to not go off and make a bunch of random changes on its own, even after being very clear not to do so, intercepting to tell it to stop doing that, then it just continues on fucking everything up.

That's my experience too, when I give Gemini CLI a big, general task and just let it run.

But if I give it structure so it can write its own context, it is truly astonishing.

I'll describe my big, general task and tell it to first read the codebase and then write a detailed requirements document, and not to change any code.

Then I'll tell it to read the codebase and the detailed requirements document it just wrote, and then write a detailed technical spec with API endpoints, params, pseudocode for tricky logic, etc.

Then I'll tell it to read the codebase, and the requirements document it just wrote, and the tech spec it just wrote, and decomp the whole development effort into weekly, daily and hourly tasks to assign to developers and save that in a dev plan document.

Only then is it ready to write code.

And I tell it to read the code base, requirements, tech spec and dev plan, all of which it authored, and implement Phase 1 of the dev plan.

It's not all mechanical and deterministic, or I could just script the whole process. Just like with a team of junior devs, I still need to review each document it writes, tweak things I don't like, or give it a better prompt to reflect my priorities that I forgot to tell it the first time, and have it redo a document from scratch.

But it produces 90% or more of its own context. It ingests all that context that it mostly authored, and then just chugs along for a long time, rarely going off the rails anymore.

Re: Gemini 2.5 Deep Think

#76

Post by golfer »

foundry27 wrote: post #34

I started doing some experimentation with this new Deep Think agent, and after five prompts I reached my daily usage limit. For $250 USD/mo that’s what you’ll be getting folks. It’s just bizarrely uncompetitive with o3-pro and Grok 4 Heavy. Anecdotally (from my experience) this was the one feature that enthusiasts in the AI community were interested in to justify the exorbitant price of Google’s Ultra subscription. I…

What was the experimentation? Can you share with us so we can see how "bizarrely uncompetitive" it is?

Re: Gemini 2.5 Deep Think

#77

Post by masterjack »

mettamage wrote: post #71
littlestymaar wrote: post #37

Earlier quoted context omitted.

That this kind of approach works is good news for local LLM enthusiasts, as it makes Cloud LLM using this more expensive while local LLM can do so for free up to a point (because LLM inference is limited by memory bandwidth not compute, you can run multiple queries in parallel on your graphic card at the same speed as the single one. Until you become compute-bound of course).

Wait, how does this work? If you load in one LLM of 40 GB, then to load in four more LLMs of 40 GB still takes up an extra 160 GB of memory right?

It will typically be the same 40 GB model loaded in, but called with many different inputs simultaneously

Re: Gemini 2.5 Deep Think

#78

Post by dweekly »

amelius wrote: post #67
foundry27 wrote: post #34

I started doing some experimentation with this new Deep Think agent, and after five prompts I reached my daily usage limit. For $250 USD/mo that’s what you’ll be getting folks. It’s just bizarrely uncompetitive with o3-pro and Grok 4 Heavy. Anecdotally (from my experience) this was the one feature that enthusiasts in the AI community were interested in to justify the exorbitant price of Google’s Ultra subscription. I…

It could be that your problem was too simple to justify the use of Deep Think. But yes, Google should have figured that out and used a less expensive mode of reasoning.

"I'm sorry but that wasn't a very interesting question you just asked. I'll spare you the credit and have a cheaper model answer that for you for free. Come back when you have something actually challenging."

Re: Gemini 2.5 Deep Think

#79

Post by irthomasthomas »

barapa wrote: post #68
irthomasthomas wrote: post #40

You can spin up a version of this at home using simonw's LLM cli with the llm-consortium plugin. Bonus 1: Use any combination of models. Mix n match models from any lab. Bonus 2: Serve your custom consortium on a local API from a single command using the llm-model-gateway plugin and use it in your apps and coding assistants. https://x.com/karpathy/status/1870692546969735361 > uv tool install llm llm install llm-conso…

I am not seeing this llm serve command

it's a separate plugin rn. llm install llm-model-gateway

Re: Gemini 2.5 Deep Think

#80

Post by breppp »

baggachipz wrote: post #28
serialNumber wrote: post #25

Earlier quoted context omitted.

I’m never the one to defend AI, but what do you mean? Is it the “AI overview” that pops up on Google? Other than that, I would say Gemini is definitely less in your face than ChatGPT for example

My company uses google workspace and every google doc, spreadsheet, calendar, online meeting and search puts nonstop callouts and messages about using Gemini. It's gotten so bad that I'm about to try building a browser extension to block that bullshit. It clutters the UI and nags. If I wanted that crap, I'd turn it on.

Yes it smells like google+, but at least gemini is actually a good product
Post reply on HN