Comment by cainxinth
1 day ago
It's the same story every time OpenAI or Anthropic releases a new model. They are generous with compute for the first few days, and use maximum fidelity with uncompressed weights. Everything runs at its best to make a good first impression. But eventually they pare things back and the models perform a little worse.
The most charitable explanation I can think of for this is something like regression to the mean. When a model is first released, there'll be a subset of users who, just by chance, sample the highest quality band of the distribution that answers their query. Some of them will rush over to social media and post about how amazing a model is. Over time, those users' mental model of responses will converge but they'll perceive the model's return to typical performance as a downgrade.
This guess/explanation predicts that most users won't match what the initial social media hype claims, doesn't discount user experience as simple habituation nor does it assume companies are lying when they say there have been no changes to the model itself (quantization included).
I also think there's an aspect where initial testing is more forgiving because the more persnickety polish bits can be ignored and tests are likely to have similar structure to things that can be trained for. Meanwhile, actual specific work items are a broader unusual distribution with more stringent acceptance criteria.
Personally, I can detect a separation between Sol and Astra (but not as large as that between Opus and Fable). While they can solve most of the same problems, Astra takes less time, is less frustrating to talk to, is cleaner, notices more, spins wheels less and requires less corrections.
> I also think there's an aspect where initial testing is more forgiving because the more persnickety polish bits can be ignored and tests are likely to have similar structure to things that can be trained for. Meanwhile, actual specific work items are a broader unusual distribution with more stringent acceptance criteria.
IMO this is 90% of it (as someone who has a bit of a different interaction style and runs these things less autonomously, and hasn't generally seen the claimed regressions). Day 1: throw new stuff at it that failed badly, exciting to see something make more progress! Day n: reality sets in that it still wasn't perfect the first time.
To add onto this, if you use a shiny new model and it gives you a turd, you're not going to tweet about it ("hey guys, look what I made with Astra! Nothing!"), and even if you do nobody is going to interact with it so it does poorly in the algorithm, because it has to compete with all the people using the new model to make something that looks impressive. Then people get tired of the magic trick and the logic flips.
Really? there would be complaints, it’s expensive and doesn’t do as well
1 reply →
We've seen that some---gpt5 was considered pretty lackluster intially, in particular. Opus 4.7 and 5 vs 4.6 were also greeted with a lot more "meh" than 4.6 or Fable.
This is a bit of an urban myth. There are trackers which keep historical performance and Sol hasn't been nerfed: https://marginlab.ai/trackers/codex/
It’s all speculation (you too), but I think the effect you’re describing is instead getting calibrated to the model’s limits. Next time a new model comes out, wait a month before trying and see if you have the same feeling of rapid quality decline after a few days. I did after I jumped back into it mid 5.x or whatever ChatGPT after paternity leave. Blown away for a few days, worried about my job for a few days, then increasingly aware of its limits.
You think they introduce stronger quantization after a few days?
For sure they quickly move to q8, the output quality difference to bf16 is small compared to the speed/capacity gain
Yeah q8 made so littler difference back when I was testing such things I'd be surprised if people could quickly notice that as a change. It's got to be either further quantized or some other type of optimization that kicks in when people notice the drop.
3 replies →
Some people here have remarked previously that while reduced precision doesn't show up in quick prompts, it does severely impact these models' ability to perform long running tasks - to the point that running these big models with severe quantization might be counterproductive as smaller but less quantized ones perform better.
Could this explain Opus?
Or a lot worse
so, AGI is cancelled?
AGI for the peasants is cancelled.
1 reply →
Do you have hard evidence of this assertion?
We can't have hard evidence. It's a SaaS and they own the code and the machine it runs on.
So it may be a widespread hallucination. But there's no evidence of that either.
Run a benchmark with a large number of samples, rerun a few days later. Compare results, use statistics to see if there's a statistically significant difference.
3 replies →
We could still have soft evidence though. Make a Todo app on Monday, and make a Todo app on Tuesday, and see what it makes in comparison.
4 replies →
There is a toot from an Open AI person a couple days ago saying they are "pulling all the levers" because of capacity issues. I have no idea what the heck the person is talking about, but I'm guessing there are consequence for those levers.
Depends on the nature of the levers
I am sure every input send to openai is prechecked by a dumb model and then send to another one. They heavily tweak this to improve performance.