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

Gemini 3

blog.google

Post reply on HN
Print view View original

861–870 of 1001 posts

Re: Gemini 3

#861

Post by egeozcan »

falcor84 wrote: post #579

I love it that there's a "Read AI-generated summary" button on their post about their new AI. I can only expect that the next step is something like "Have your AI read our AI's auto-generated summary", and so forth until we are all the way at Douglas Adams's Electric Monk: > The Electric Monk was a labour-saving device, like a dishwasher or a video recorder. Dishwashers washed tedious dishes for you, thus saving you…

I'm afraid they will finish "The Salmon of Doubt" with AI and sell it to the future generations with a very small disclaimer, stating it's inspired by Douglas Adams.

The possibility was already a topic in the series "Mozart in the jungle" where they made a robot which supposedly finished the Requiem piece by Mozart.

Re: Gemini 3

#862

Post by egeozcan »

stephc_int13 wrote: post #549
tylervigen wrote: post #313

Earlier quoted context omitted.

I agree that for any given test, you could build a specific pipeline to optimize for that test. I supposed that's why it is helpful to have many tests. However, many people have worked hard to optimize tools specifically for ARC over many years, and it's proven to be a particularly hard test to optimize for. This is why I find it so interesting that LLMs can do it well at all, regardless of whether tests like it are…

The real strength of current neural nets/transformers relies on huge datasets. ARC do not provide this kind of dataset, only a small public one and a private one where they do the benchmarks. Building your own large private ARC set does not seem too difficult if you have enough resources.

How can they keep it private? It's not like they can run these models locally. Do the providers promise not to peak when they are testing?

Re: Gemini 3

#863

Post by column »

Workaccount2 wrote: post #100

It still failed my image identification test ([a photoshopped picture of a dog with 5 legs]...please count the legs) that so far every other model has failed agonizingly, even failing when I tell them they are failing, and they tend to fight back at me. Gemini 3 however, while still failing, at least recognized the 5th leg, but thought the dog was...well endowed. The 5th leg however is clearly a leg, despite being wh…

"[a photoshopped picture of a dog with 5 legs]...please count the legs"

Meanwhile you could benchmark for something actually useful. If you're about to say "But that means it won't work for my use case of identifying a person on a live feed" or whatever, then why don't you test that? I really don't understand the kick people get of successfully tricking LLMs on non productive task with no real world application. Just like the "how many r in strawberry?", "uh uh uh it says two urh urh".. ok but so what? What good is a benchmark that is so far from a real use case?

Re: Gemini 3

#864

Post by hollowturtle »

Kiro wrote: post #832
hollowturtle wrote: post #778

Earlier quoted context omitted.

Wow looks like total shit and eventually very hard to take on and actually improve it, given the convoluted code it generated, YET people are impressed. What world are we living in...

You can criticize the code but "wow looks like total shit" is such an embarrassing thing to say considering the context. Imagine going back a few years and show them a tool outputting this from text. No-one would believe it.

It simply is non impressive at all to me, we had an industry(games not web) that was the most innovativd and was able to do things, and in part still is, thousands of years ahead of the slop glorified here

Re: Gemini 3

#865

Post by hollowturtle »

SXX wrote: post #810
hollowturtle wrote: post #778

Earlier quoted context omitted.

Wow looks like total shit and eventually very hard to take on and actually improve it, given the convoluted code it generated, YET people are impressed. What world are we living in...

You are missing the point of this exercise. This is not about code quality - its about capacity of model to generate visuals with no guidance. For the code quality it can really be as good or as bad ad as you desire. In this case it is what it is because I put zero effort into it.

non impressive at all to me, visuals are bad not even a student starting in animations would produce that slop. You're glorifying slop, as for the code quality that's not about styling or semantics the tecniques used are BAD and won't scale at all, eg setTimeout is not designed to be run at exactly that interval, it's just a timeout suggestion. And no it cannot be good or bad as you desire it's just bad, I have YET to see something better than an animation student on the first year would do. You're destroying the software industry with this mentality

Re: Gemini 3

#866

Post by pk-protect-ai »

It is pointless to ask an LLM to draw an ASCII unicorn these days. Gemini 3 draws one of these (depending on the prompt):

https://www.ascii-art.de/ascii/uvw/unicorn.txt

However, it is amazing how far spatial comprehension has improved in multimodal models.

I'm not sure the below would be properly displayed on HN; you'll probably need to cut and paste it into a text editor.

Prompt: Draw me an ASCII world map with tags or markings for the areas and special places.

Temperature: 1.85

Top-P 0.98

Answer: Edit (replaced with URL) https://justpaste.it/kpow3

Re: Gemini 3

#867

Post by nprateem »

OMG they've obviously had a major breakthrough because now it can reply to questions with actual answers instead of shit blog posts.

Re: Gemini 3

#868

Post by cafebabbe »

wraptile wrote: post #796
ogig wrote: post #711

I just gave it a short description of a small game I had an idea for. It was 7 sentences. It pretty much nailed a working prototype, using React, clean css, Typescript and state management. It event implemented a Gemini query using the API for strategic analysis given a game state. I'm more than impressed, I'm terrified. Seriously thinking of a career change.

I find it funny to find this almost exact same post in every new model release thread. Yet here we are - spending the same amount of time, if not more, finishing the rest of the owl.

Seems like the whole world forgot what this job was really about :/

Re: Gemini 3

#869

Post by thomasahle »

panarky wrote: post #697
lairv wrote: post #293

Out of curiosity, I gave it the latest project euler problem published on 11/16/2025, very likely out of the training data Gemini thought for 5m10s before giving me a python snippet that produced the correct answer. The leaderboard says that the 3 fastest human to solve this problem took 14min, 20min and 1h14min respectively Even thought I expect this sort of problem to very much be in the distribution of what the mo…

[flagged]

From my understanding, Google put online the largest RL cluster in the world not so long ago. It's not surprising they do really well on things that are "easy" to RL, like math or SimpleQA

Re: Gemini 3

#870

Post by lan321 »

marcus_holmes wrote: post #822
jorvi wrote: post #775

Earlier quoted context omitted.

Same. I've been needing to update an userscript (JS) that takes stuff like "3 for the price of 1", "5 + 1 free", "35% discount!" from a particular site and then converts the price to a % discount and the price per item / 250 grams. Its an old userscript so it is glitchy and halfway works. I already pre-chewed the work by telling Gemini 3 exactly which new HTML elements it needs to match and which contents it needs to…

1. Coding with LLMs seems to be all about context management. Getting the LLM to deal with the minimum amount of code needed to fix the problem or build the feature, carefully managing token limits and artificially resetting the session when needed so the context handover is managed, all that. Just pointing an LLM at a large code base and expecting good things doesn't work. 2. I've found the same with Gemini; I can r…

The problem is by that point it's much less useful in projects. I still like them but when I get to the point of telling it exactly what to do I'm mostly just being lazy. It's useful in that it might give me some ideas I didn't consider but I'm not sure it's saving time.

Of course, for short one-off scripts, it's amazing. It's also really good at preliminary code reviews. Although if you have some awkward bits due to things outside of your power it'll always complain about them and insist they are wrong and that it can be so much easier if you just do it the naive way.

Amazon's Kiro IDE seems to have a really good flow, trying to split large projects into bite sized chunks. I, sadly, couldn't even get it to implement solitaire correctly, but the idea sounds good. Agents also seem to help a lot since it can just do things from trial and error, but company policy understandably gets complicated quick if you want to provide the entire repo to an LLM agent and run 'user approved' commands it suggests.

Post reply on HN