I've noticed something else - as Anthropic models get even more and more superhuman, they seem to serve me more and more casual nonsense.
Not like adding glue to pizza. Here's an example from today (paraphrasing): "you need to run `git merge-base branch1 branch2`. Pay attention to the order of arguments, it is important: `git merge-base` is symmetric and returns the same value regardless of the order of inputs".
So which one is it? Symmetric or not? It's not even one of those where it self-corrects, it just happily contradicts itself halfway through the sentence.
My pet theory is that these frontier models do quite a bit of brute force at the end - some kind of beam search - and silently downgrade you depending on demand or compute availability.
Still not great. This particular nugget is from Sonnet 5, default settings.
I also see stuff like this with Opus (4.8 and 5). It does the work correctly but then the explanation or "aftermath" description is kinda nonsense. It feels like they are tuned hard for agentic coding and everything else is getting worse. There are people that say older versions are also better for creative writing which would confirm this theory.
Claude has always noticeably degraded under heavy load. Opus goes from "decent to work with" to "dumb intern" depending on whether you're working at 3 AM west coast or 10 AM - 5 PM. It's part of why I cancelled my subscription - "max" plans and "extra high" effort are meaningless when there's so much variability between model availability, model performance, and harness bugs every day and every week.
I've noticed the same pattern even in GLM5.2. It has always been a thing, but it seems to be getting worse in recent models
It does feel like the kind of thing beam search would fix. The LLM starts the sentence with a claim like "Pay attention to the order of arguments". Around that time it "notices" that the order doesn't matter, but it's already committed to the sentence and has to complete it in the best way still possible
Maybe at some point someone figures out how to train models with a backspace token
Yes, LLMs are prone to semantic consistency traps. I see a version of this in a system I have that transcribes a lot of noisy low quality audio (radio comms).
Sometimes it works out, in that an unrecognizable word or two is replaced with reasonable assumptions based on the semantics established by the words that came before the signal degraded.
But the bad audio might also result in words that don't align well to what came before, or represent alternate (mis)interpretations. Now this is part of the context and the next several tokens align to this new path regardless of what is said in the audio.
In pre-LLM transcription, you might get a nonsense word or two when the audio transiently degrades, but the specific meaning of the nonsense words doesn't influence the transcription of audio following the degradation.
Have noticed similar things using it for code review. "Issue #3: This variable is defined but never used. It is actually used later in the method though. So this is not a real issue to be concerned about", that sort of thing.
First we started by extending chain-of-thought to a formal thinking system. You could extend this again with a "thought scratch-pad" or "thinking draft" with n-number of passes without needing to train backspace tokens.
That being said, all of these are just variations of scaffolding on generating more tokens.
Sounds like a way to distract from the real environmental issues with yet another "plastic straws" debacle. A fake issue that makes for some nice headlines in the media - but when you run the numbers, it obviously doesn't matter and never did.
Distract from the real issues like, you know. Clean energy projects in the US being throttled on federal level due to lobbying by the oil companies - lobbying that borders on regulatory capture. At the time when the oil prices are rising globally - and the demand for energy in the US is rising too.
Unfortunately, you can't stop an environmentalist from throwing his weight behind the cause that's being promoted the loudest. So influence groups can just push one meaningless "headline cause" after another, and keep them distracted from anything that matters in perpetuity.
Yes. I switched to Codex some time ago. I feel like I get fewer outages (had one over the weekend), and coding results are good. I built my own agent harness using OpenAI models too and feel it was the right decision.
Third outage today. Per https://downdetector.com/status/claude-ai/ the people affected grows every time: first one had peak 19 reports, second 24 and now it's already 39.
exac | 12 hours ago
rich_sasha | 10 hours ago
Not like adding glue to pizza. Here's an example from today (paraphrasing): "you need to run `git merge-base branch1 branch2`. Pay attention to the order of arguments, it is important: `git merge-base` is symmetric and returns the same value regardless of the order of inputs".
So which one is it? Symmetric or not? It's not even one of those where it self-corrects, it just happily contradicts itself halfway through the sentence.
My pet theory is that these frontier models do quite a bit of brute force at the end - some kind of beam search - and silently downgrade you depending on demand or compute availability.
Still not great. This particular nugget is from Sonnet 5, default settings.
sunaookami | 9 hours ago
Culonavirus | 8 hours ago
Well you can't blame them, it's what brings in the cash. Other uses will suffer proportionally.
o10449366 | 8 hours ago
wongarsu | 8 hours ago
It does feel like the kind of thing beam search would fix. The LLM starts the sentence with a claim like "Pay attention to the order of arguments". Around that time it "notices" that the order doesn't matter, but it's already committed to the sentence and has to complete it in the best way still possible
Maybe at some point someone figures out how to train models with a backspace token
zhoBEENG | 7 hours ago
sjsdaiuasgdia | 7 hours ago
Sometimes it works out, in that an unrecognizable word or two is replaced with reasonable assumptions based on the semantics established by the words that came before the signal degraded.
But the bad audio might also result in words that don't align well to what came before, or represent alternate (mis)interpretations. Now this is part of the context and the next several tokens align to this new path regardless of what is said in the audio.
In pre-LLM transcription, you might get a nonsense word or two when the audio transiently degrades, but the specific meaning of the nonsense words doesn't influence the transcription of audio following the degradation.
maweaver | 7 hours ago
willmadden | 7 hours ago
bellowsgulch | 3 hours ago
That being said, all of these are just variations of scaffolding on generating more tokens.
brookst | 8 hours ago
croemer | 9 hours ago
Update 11:27 UTC: I saw the error first at 11:22 UTC. Retry at 11:27 UTC still failing. Status page is still green.
Update 11:28 UTC: Incident has been declared dated 11:27 UTC https://status.claude.com/incidents/mfdtrknpxghq
Khaine | 9 hours ago
rvz | 9 hours ago
No wonder the amount of water that both Claude and Codex are taking they also need so many frequent hydration breaks.
[0] https://news.ycombinator.com/item?id=49056739
ModernMech | 9 hours ago
ACCount37 | 9 hours ago
rvz | 8 hours ago
So you think that the water that comes out of these data centers is safe for humans once released and the mass consumption of them is not a concern?
Sounds like a way to sweep this environmental issue under the rug.
[0] https://theoec.org/news-and-information/behind-the-data-boom...
[1] https://fieldreport.caes.uga.edu/publications/TP121/how-data...
[2] https://www.wsj.com/tech/ai/ai-data-centers-water-use-901e29...
ACCount37 | 4 hours ago
Distract from the real issues like, you know. Clean energy projects in the US being throttled on federal level due to lobbying by the oil companies - lobbying that borders on regulatory capture. At the time when the oil prices are rising globally - and the demand for energy in the US is rising too.
Unfortunately, you can't stop an environmentalist from throwing his weight behind the cause that's being promoted the loudest. So influence groups can just push one meaningless "headline cause" after another, and keep them distracted from anything that matters in perpetuity.
breezybottom | 7 hours ago
matheusmoreira | 8 hours ago
seunosewa | 8 hours ago
vintagedave | 8 hours ago
croemer | 7 hours ago
Third outage today. Per https://downdetector.com/status/claude-ai/ the people affected grows every time: first one had peak 19 reports, second 24 and now it's already 39.