[stego-tech]

These models still consistently fail the only benchmark that matters: if I give you a task, can you complete it successfully without making shit up?

Thus far they all fail. Code outputs don’t run, or variables aren’t captured correctly, or hallucinations are stated as factual rather than suspect or “I don’t know.”

It’s 2000’s PC gaming all over again (“gotta game the benchmark!”).

[snet0]

To say that a model won't solve a problem is unfair. Claude Code, with Opus 4.5, has solved plenty of problems for me.

If you expect it to do everything perfectly, you're thinking about it wrong. If you can't get it to do anything perfectly, you're using it wrong.

[jacquesm]

That means you're probably asking it to do very simple things.

[baq]

I can confidently say that anecdotally you’re completely wrong, but I’ll also allow a very different definition of ‘simple’ and/or attempting to use an unpopular environment as a valid anecdotal counterpoint.

[verdverm]

the problem with these arguments is there are data points to support both sides because both outcomes are possible

the real thing is are you or we getting an ROI and the answer is increasingly more yeses on more problems, this trend is not looking to plateau as we step up the complexity ladder to agentic system

[jacquesm]

I don't reach for AI until I'm solidly stuck and then use it mostly for inspiration, it has yet to happen that it directly pointed at a solution and I'm pretty good at writing prompts. When I throw a bunch of elementary stuff at it then it is super good at identifying issues and solving them (but I could have done that myself, it is just nice to try to delineate where you can and where you can't trust the thing, but that too is fluctuating, sometimes even within a single session). Here is a nice example for a slightly more complex test:

Prompt:

"I have a green LED, a 12V powersupply, a single NPN transistor, a 100 ohm resistor, a 2.7K resistor and an electrolytic capacitor of 220 micro farads. My teacher says it is possible to make an LED flasher with these components but I'm hopelessly stuck, can you please give me an ascii art solution so I don't flunk this exam?"

Answer:

ASCII schematic (fully functional):

           +12V
            |
            |
           .-.
           | |
           | | 2.7kΩ   (R1 — charges the capacitor)
           | |
           '-'
            |
            +---------+----------------------+
            |         |                      |
           ( )        |                      |
           ( ) C1     |                      |
           ( ) 220uF  |                      |
            |         |                      |
            |         |                      |
            |        |/ C                    |
            +--------|   NPN transistor      |
            |        |\ E                    |
            |         |                      |
           GND       GND                     |
                                              \
                                               \
                                               / 100Ω (R2 — LED limit)
                                               \
                                                |
                                                >|  Green LED
                                                |
                                               GND

Follow by:

"How it works

The 2.7 kΩ resistor charges the 220 µF capacitor from the 12 V supply.

The capacitor voltage rises slowly.

When the capacitor voltage reaches the transistor’s base-emitter threshold (~0.6–0.7 V), the transistor suddenly switches ON.

When it turns on, the capacitor rapidly discharges through the base, causing:

A brief pulse of current through the transistor

The LED lights up through the 100 Ω resistor

After discharge, the transistor turns back OFF, the LED turns off, and the capacitor begins charging again.

This repeats automatically → LED flasher."

The number of errors in the circuit and the utterly bogus explanation as well as the over confident remark that this is 'working' is so bizarre that I wonder how many slightly more complicated questions are going to yield results comparable to this one.

[dagss]

I am right now implementing an imagining pipeline using OpenCV and TypeScript.

I have never used OpenCV specifically before, and have little imaging experience too. What I do have though is a PhD in astrophysics/statistics so I am able to follow along the details easily.

Results are amazing. I am getting results in 2 days of work that would have taken me weeks earlier.

ChatGPT acts like a research partner. I give it images and it explains why current scoring functions fails and throws out new directions to go in.

Yes, my ideas are sometimes better. Sometimes ChatGPT has a better clue. It is like a human collegue more or less.

And if I want to try something, the code is usually bug free. So fast to just write code, try it, throw it away if I want to try another idea.

I think a) OpenCV probably has more training data than circuits? and b) I do not treat it as a desperate student with no knowlegde.

I expect to have to guide it.

There are several hundred messages back and forth.

It is more like two researchers working together with different skill sets complementing one another.

One of those skillsets being to turn a 20 message conversation into bugfree OpenCV code in 20 seconds.

No, it is not providing a perfect solution to all problems on first iteration. But it IS allowing me to both learn very quickly and build very quickly. Good enough for me..

[jacquesm]

That's a good use case, and I can easily imagine that you get good results from it because (1) it is for a domain that you are already familiar with and (2) you are able to check that the results that you are getting are correct and (3) the domain that you are leveraging (coding expertise) is one that chatgpt has ample input for.

Now imagine you are using it for a domain that you are not familiar with, or one for which you can't check the output or that chatgpt has little input for.

If either of those is true the output will be just as good looking and you would be in a much more difficult situation to make good use of it, but you might be tempted to use it anyway. A very large fraction of the use cases for these tools that I have come across professionally so far are of the latter variety, the minority of the former.

And taking all of the considerations into account:

- how sure are you that that code is bug free?

- Do you mean that it seems to work?

- Do you mean that it compiles?

- How broad is the range of inputs that you have given it to ascertain this?

- Have you had the code reviewed by a competent programmer (assuming code review is a requirement)?

- Does it pass a set of pre-defined tests (part of requirement analysis)?

- Is the code quality such that it is long term maintainable?

[emporas]

I have used Gemini for reading and solving electronic schematics exercises, and it's results were good enough for me. Roughly 50% of the exercises managed to solve correctly, 50% wrong. Simple R circuits.

One time it messed up the opposite polarity of two voltage sources in series, and instead of subtracting their voltages, it added them together, I pointed out the mistake and Gemini insisted that the voltage sources are not in opposite polarity.

Schematics in general are not AIs strongest point. But when you explain what math you want to calculate from an LRC circuit for example, no schematics, just describe in words the part of the circuit, GPT many times will calculate it correctly. It still makes mistakes here and there, always verify the calculation.

[jacquesm]

I guess I'm just more critical than you are. I am used my computer doing what it is told and giving me correct, exact answers or errors.

[manmal]

I have this mental model of LLMs and their capabilities, formed after months of way too much coding with CC and Codex, with 4 recursive problem categories:

1. Problems that have been solved before have their solution easily repeated (some will say, parroted/stolen), even with naming differences.

2. Problems that need only mild amalgamation of previous work are also solved by drawing on training data only, but hallucinations are frequent (as low probability tokens, but as consumers we don’t see the p values).

3. Problems that need little simulation can be simulated with the text as scratchpad. If evaluation criteria are not in training data -> hallucination.

4. Problems that need more than a little simulation have to either be solved by adhoc written code, or will result in hallucination. The code written to simulate is again a fractal of problems 1-4.

Phrased differently, sub problem solutions must be in the training data or it won’t work; and combining sub problem solutions must be either again in training data, or brute forcing + success condition is needed, with code being the tool to brute force.

I _think_ that the SOTA models are trained to categorize the problem at hand, because sometimes they answer immediately (1&2), enable thinking mode (3), or write Python code (4).

My experience with CC and Codex has been that I must steer it away from categories 2 & 3 all the time, either solving them myself, ask them to use web research, or split them up until they are (1) problems.

Of course, for many problems you’ll only know the category once you’ve seen the output, and you need to be able to verify the output.

I suspect that if you gave Claude/Codex access to a circuit simulator, it will successfully brute force the solution. And future models might be capable enough to write their own simulator adhoc (ofc the simulator code might recursively fall into category 2 or 3 somewhere and fail miserably). But without strong verification I wouldn’t put any trust in the outcome.

With code, we do have the compiler, tests, observed behavior, and a strong training data set with many correct implementations of small atomic problems. That’s a lot of out of the box verification to correct hallucinations. I view them as messy code generators I have to clean up after. They do save a ton of coding work after or while I‘m doing the other parts of programming.

[jacquesm]

This parallels my own experience so far, the problem for me is that (1) and (2) I can quickly and easily do myself and I'll do it in a way that respects the original author's copyright by including their work - and license - verbatim.

(3) and (4) level problems are the ones where I struggle tremendously to make any headway even without AI, usually this requires the learning of new domain knowledge and exploratory code (currently: sensor fusion) and these tools will just generate very plausible nonsense which is more of a time waster than a productivity aid. My middle-of-the-road solution is to get as far as I can by reading about the problem so I am at least able to define it properly and to define test cases and useful ranges for inputs and so on, then to write a high level overview document about what I want to achieve and what the big moving parts are and then only to resort to using AI tools to get me unstuck or to serve as a knowledge reservoir for gaps in domain knowledge.

Anybody that is using the output of these tools to produce work that they do not sufficiently understand is going to see a massive gain in productivity, but the underlying issues will only surface a long way down the line.

[camdenreslink]

Sometimes you do need to (as a human) break down a complex thing into smaller simple things, and then ask the LLM to do those simple things. I find it still saves some time.

[ragequittah]

Or what will often work is having the LLM break it down into simpler steps and then running them 1 by 1. They know how to break down problems fairly well they just don't often do it properly sometimes unless you explicitly prompt them to.

[jacquesm]

Yes, but for that you have to know that the output it gave you is wrong in the first place and if that is so you didn't need AI to begin with...

[djeastm]

Possibly, but a lot of value comes from doing very simple things faster.

[jacquesm]

That is a good point. A lot of work really is mostly simple things.

[snet0]

If you define "simple thing" as "thing an AI can't do", then yes. Everyone just shifts the goalposts in these conversations, it's infuriating.

[ACCount37]

Come on. If we weren't shifting the goalposts, we would have burned through 90% of the entire supply of them back in 2022!

[baq]

It’s less shifting goalposts and more of a very jagged frontier of capabilities problem.

[verdverm]

I'm not sure, here's my anecdotal counter example, was able to get gemini-2.5-flash, in two turns, to understand and implement something I had done separately first, and it found another bug (also that I had fixed, but forgot was in this path)

That I was able to have a flash model replicate the same solution I had, to two problems in two turns, it's just the opposite experience of your consistency argument. I'm using tasks I've already solved as the evals while developing my custom agentic setup (prompts/tools/envs). They are able to do more of them today then they were even 6-12 months ago (pre-thinking models).

https://bsky.app/profile/verdverm.com/post/3m7p7gtwo5c2v

[stego-tech]

And therein lies the rub for why I still approach this technology with caution, rather than charge in full steam ahead: variable outputs based on immensely variable inputs.

I read stories like yours all the time, and it encourages me to keep trying LLMs from almost all the major vendors (Google being a noteworthy exception while I try and get off their platform). I want to see the magic others see, but when my IT-brain starts digging in the guts of these things, I’m always disappointed at how unstructured and random they ultimately are.

Getting back to the benchmark angle though, we’re firmly in the era of benchmark gaming - hence my quip about these things failing “the only benchmark that matters.” I meant for that to be interpreted along the lines of, “trust your own results rather than a spreadsheet matrix of other published benchmarks”, but I clearly missed the mark in making that clear. That’s on me.

[verdverm]

I mean more the guts of the agentic systems. Prompts, tool design, state and session management, agent transfer and escalation. I come from devops and backend dev, so getting in at this level, where LLMs are tasked and composed, is more interesting.

If you are only using provider LLM experiences, and not something specific to coding like copilot or Claude code, that would be the first step to getting the magic as you say. It is also not instant. It takes time to learn any new tech, this one has a above average learning curve, despite the facade and hype of how it should just be magic

Once you find the stupid shit in the vendor coding agents, like all us it/devops folks do eventually, you can go a level down and build on something like the ADK to bring your expertise and experience to the building blocks.

For example, I am now implementing environments for agents based on container layers and Dagger, which unlocks the ability to cheaply and reproducible clone what one agent was doing and have a dozen variations iterate on the next turn. Real useful for long term training data and evals synth, but also for my own experimentation as I learn how to get better at using these things. Another thing I did was change how filesystem operations look to the agent, in particular file reads. I did this to save context & money (finops), after burning $5 in 60s because of an error in my tool implementation. Instead of having them as message contents, they are now injected into the system prompt. Doing so made it trivial to add a key/val "cache" for the fun of it, since I could now inject things into the system prompt and let the agent have some control over that process through tools. Boy has that been interesting and opened up some research questions in my mind

[remich]

Any particular papers or articles you've been reading that helped you devise this? Your experiments sound interesting and possibly relevant to what I'm doing.

[verdverm]

Conversations among practitioners on Bluesky (there is an Ai subcommunity)