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

Gemini 2.5 Pro Preview

developers.googleblog.com

Post reply on HN
Print view View original

611–620 of 728 posts

Re: Gemini 2.5 Pro Preview

#611

Post by kupopuffs »

thewebguyd wrote: post #389
mattlondon wrote: post #308

Earlier quoted context omitted.

AIs aren't intended to be used as calculators though? You could say that when I use my spanner/wrench to tighten a nut it works 100% of the time, but as soon as I try to use a screwdriver it's terrible and full of problems and it can't even reliably so something as trivially easy as tighten a nut, even though a screwdriver works the same way by using torque to tighten a fastener. Well that's because one tool is desig…

> AIs aren't intended to be used as calculators though? Then why are we using them to write code, which should produce reliable outputs for a given input...much like a calculator. Obviously we want the code to produce correct results for whatever input we give, and as it stands now, I can't trust LLM output without reviewing first. Still a helpful tool, but ultimately my desire would be to have them be as accurate as…

I dunno man, I think writing an app is 10000x harder than adding 5 + 5

Re: Gemini 2.5 Pro Preview

#612

Post by namaria »

mark_l_watson wrote: post #479
gtirloni wrote: post #478

Earlier quoted context omitted.

That's your extraordinary evidence?

Nope, just my opinion, derived from watching monthly and weekly exponential improvement over a few year period. I worked through at least two,AI winters since 1982, so current progress is good to see.

Exponential over which metric exactly? Training dataset size, compute required yeah these have grown exponentially. But has any measure capability?

Because exponentially growing costs with linear or not measurable improvements is not a great trajectory.

Re: Gemini 2.5 Pro Preview

#613

Post by floydnoel »

Jordan-117 wrote: post #247
segphault wrote: post #18

My frustration with using these models for programming in the past has largely been around their tendency to hallucinate APIs that simply don't exist. The Gemini 2.5 models, both pro and flash, seem significantly less susceptible to this than any other model I've tried. There are still significant limitations, no amount of prompting will get current models to approach abstraction and architecture the way a person doe…

I recently needed to recommend some IAM permissions for an assistant on a hobby project; not complete access but just enough to do what was required. Was rusty with the console and didn't have direct access to it at the time, but figured it was a solid use case for LLMs since AWS is so ubiquitous and well-documented. I actually queried 4o, 3.7 Sonnet, and Gemini 2.5 for recommendations, stripped the list of duplicate…

by asking three different models and then keeping everything single unique thing they gave you, i believe you actually maximized your chances of running into hallucinations.

instead of ignoring the duplicates, when i query different models, i use the duplicates as a signal that something might be more accurate. i wonder what your results might have looked like if you only kept the duplicated permissions and went from there.

Re: Gemini 2.5 Pro Preview

#614

Post by lopatin »

minzi wrote: post #610

I use Gemini inside cursor, but the web app is basically unusable to me. Of the big three, only Claude seems to have a sensible web app with good markdown formatting, converting big pastes into attachments, and breaking out code into side panels. These seem like relatively obvious features so it’s confusing to me that Google is so behind on the UI here.

My biggest issue with the UI is that it doesn’t let you edit your prompt if it’s not the latest one. If someone at Google is reading this, can you just implement it lol. Such a minor thing but really disrupts my workflow when dealing with a long conversation. I want to “rewind” the conversation basically.

Re: Gemini 2.5 Pro Preview

#615

Post by sweezyjeezy »

namaria wrote: post #609
sweezyjeezy wrote: post #421

Earlier quoted context omitted.

I would argue that what LLMs are capable of doing right now is already pretty extraordinary, and would fulfil your extraordinary evidence request. To turn it on its head - given the rather astonishing success of the recent LLM training approaches, what evidence do you have that these models are going to plateau short of your own abilities?

On Limitations of the Transformer Architecture https://arxiv.org/abs/2402.08164 Theoretical limitations of multi-layer Transformer https://arxiv.org/abs/2412.02975

Only skimmed, but both seem to be referring to what transformers can do in a single forward pass, reasoning models would clearly be a way around that limitation.

o4 has no problem with the examples of the first paper (appendix A). You can see its reasoning here is also sound: https://chatgpt.com/share/681b468c-3e80-8002-bafe-279bbe9e18.... Not conclusive unfortunately since this is in date-range of its training data. Reasoning models killed off a large class of "easy logic errors" people discovered from the earlier generations though.

Re: Gemini 2.5 Pro Preview

#616

Post by namaria »

joshjob42 wrote: post #413
jstummbillig wrote: post #314

Earlier quoted context omitted.

> no amount of prompting will get current models to approach abstraction and architecture the way a person does I find this sentiment increasingly worrisome. It's entirely clear that every last human will be beaten on code design in the upcoming years (I am not going to argue if it's 1 or 5 years away, who cares?) I wished people would just stop holding on to what amounts to nothing, and think and talk more about wha…

I mean, if you draw the scaling curves out and believe them, then sometime in the next 3-10 years, plausibly shorter, AIs will be able to achieve best-case human performance in everything able to be done with a computer and do it at 10-1000x less cost than a human, and shortly thereafter robots will be able to do something similar (though with a smaller delta in cost) for physical labor, and then shortly after that w…

"Extrapolation" https://xkcd.com/605/

Re: Gemini 2.5 Pro Preview

#617

Post by lallysingh »

ssalazar wrote: post #454
jstummbillig wrote: post #314

Earlier quoted context omitted.

> no amount of prompting will get current models to approach abstraction and architecture the way a person does I find this sentiment increasingly worrisome. It's entirely clear that every last human will be beaten on code design in the upcoming years (I am not going to argue if it's 1 or 5 years away, who cares?) I wished people would just stop holding on to what amounts to nothing, and think and talk more about wha…

I code with multiple LLMs every day and build products that use LLM tech under the hood. I dont think we're anywhere near LLMs being good at code design. Existing models make _tons_ of basic mistakes and require supervision even for relatively simple coding tasks in popular languages, and its worse for languages and frameworks that are less represented in public sources of training data. I am _frequently_ having to t…

The software tool takes a higher-level input to produce the executable.

I'm waiting for LLMs to integrate directly into programming languages.

The discussions sound a bit like the early days of when compilers started coming out, and people had been using direct assembler before. And then decades after, when people complained about compiler bugs and poor optimizers.

Re: Gemini 2.5 Pro Preview

#618

Post by postexitus »

ArthurStacks wrote: post #600
postexitus wrote: post #597

Earlier quoted context omitted.

and you happened to have created an account in hackernews just 3 months ago after 30 years in business just to hunt AI-sceptics?

I dont hunt 'AI skeptics'. I just provide a viewpoint based on professional experience. Not one that is 'AI is bad at coding because everyone on Twitter says so"

and you happened to have created an account in hackernews just 3 months ago after 30 years in business just to provide a viewpoint based on professional experience?

Re: Gemini 2.5 Pro Preview

#619

Post by askl »

acedTrex wrote: post #358
jstummbillig wrote: post #314

Earlier quoted context omitted.

> no amount of prompting will get current models to approach abstraction and architecture the way a person does I find this sentiment increasingly worrisome. It's entirely clear that every last human will be beaten on code design in the upcoming years (I am not going to argue if it's 1 or 5 years away, who cares?) I wished people would just stop holding on to what amounts to nothing, and think and talk more about wha…

> It's entirely clear that every last human will be beaten on code design in the upcoming years In what world is this statement remotely true.

In the delusional startup world.

Re: Gemini 2.5 Pro Preview

#620

Post by concats »

jstummbillig wrote: post #314
segphault wrote: post #18

My frustration with using these models for programming in the past has largely been around their tendency to hallucinate APIs that simply don't exist. The Gemini 2.5 models, both pro and flash, seem significantly less susceptible to this than any other model I've tried. There are still significant limitations, no amount of prompting will get current models to approach abstraction and architecture the way a person doe…

> no amount of prompting will get current models to approach abstraction and architecture the way a person does I find this sentiment increasingly worrisome. It's entirely clear that every last human will be beaten on code design in the upcoming years (I am not going to argue if it's 1 or 5 years away, who cares?) I wished people would just stop holding on to what amounts to nothing, and think and talk more about wha…

I won't deny that in a context with perfect information, a future LLM will most likely produce flawless code. I too believe that is inevitable.

However, in real life work situations, that 'perfect information' prerequisite will be a big hurdle I think. Design can depend on any number of vague agreements and lots of domain specific knowledge, things a senior software architect has only learnt because they've been at the company for a long time. It will be very hard for a LLM to take all the correct decisions without that knowledge.

Sure, if you write down a summary of each and every meeting you've attended for the past 12 months, as well as attach your entire company confluence, into the prompt, perhaps then the LLM can design the right architecture. But is that realistic?

More likely I think the human will do the initial design and specification documents, with the aforementioned things in mind, and then the LLM can do the rest of the coding.

Not because it would have been technically impossible for the LLM to do the code design, but because it would have been practically impossible to craft the correct prompt that would have given the desired result from a blank sheet.

Post reply on HN