Seems insane. It isn’t that had to use anthropic managed agents to build something similar - work with an in house version that I think is superior. I’m sure Devin is a good fit for some companies…but struggling to see…
I want an e-reader that can handle narration. Actually immersive reading - https://storyteller-platform.dev - without paying Amazon. This is possible on a kindle running KOReader, but it's very slow loading for large…
Thanks for mentioning this. It makes me feel better about stumbling with metal math in front of some colleagues that are faster.
You would not expect the developers of the model to optimize for a well known benchmark?
I have a feeling that if they already thought like that, they wouldn’t be using IBM legacy stacks in 2026.
There is so much business software out there and ecosystems that have evolved around it. Systems still running cobol on IBM mainframes; or, more "modern" stacks like Java apps running in IBM websphere. There are…
I agree about the shepherding role. But, I'm not sure about how hands-off the shepherding will be at well-run companies. I work on pricing software for a regulated industry. If I were to start over again from scratch on…
looks like it worked :)
“ Name another service that does what they do for FREE (and no ads) at this scale” and is reliable is the question
Hopefully because they’re focusing on having relevance in the future. It’s sad as a gen Xer to see Japanese electronics companies fading away.
Same here. Same with not downing the harshest rye whiskey again. Executive function something or other …
It isn’t racist to acknowledge that some schools select students based on factors other than demonstrated aptitude. In the last ten years it was a fad in the US to not require SAT scores until folks realized this was…
Maybe I’m missing something. How is codex desktop more feature rich that Claude desktop? Aren’t they about the same now? I actually found codex desktop cloud mode to be more difficult to use in the codex desktop vs…
Ive been using 5.6 sol xhigh everyday since it launched. It does feel slower than it used to be. Fast mode seems useless now - burns through tokens without a noticeable speed gain for me now.
"So we are going back to blue books..." exactly. And, it is also efficient. You just need a proctor to ID the test taker and make sure they aren't using AI to cheat in order to verify that the student learned something…
charge more for it. its fine if free customers dont have service. dont break paying customers
gitlab self hosted though? no outage. i remember equivalent functionality at my last job taking this path
curious about methodology. ive seen them post results for claude/codex when they only ran over benchmarks 3 times per model...
how hard could it be these days for a mid to large size eng company to have their own gitlab/hub type solution hosted in aws
#0-nines
are you sure its actually two nines and not zero?
WTF. Outage every month if not multiple. Even for paying enterprise plans. Also, doesn't even have RAG offering.
Well then, the credible, perv glasses wearing authority on AI has spoken
Re the part about the brain associating how you feel at the moment / how you think you are performing with how you judge yourself as a person: I recommend reading Steve Hayes work / ACT therapy including this video for…
Sounds like you need to switch jobs.
Seems insane. It isn’t that had to use anthropic managed agents to build something similar - work with an in house version that I think is superior. I’m sure Devin is a good fit for some companies…but struggling to see…
I want an e-reader that can handle narration. Actually immersive reading - https://storyteller-platform.dev - without paying Amazon. This is possible on a kindle running KOReader, but it's very slow loading for large…
Thanks for mentioning this. It makes me feel better about stumbling with metal math in front of some colleagues that are faster.
You would not expect the developers of the model to optimize for a well known benchmark?
I have a feeling that if they already thought like that, they wouldn’t be using IBM legacy stacks in 2026.
There is so much business software out there and ecosystems that have evolved around it. Systems still running cobol on IBM mainframes; or, more "modern" stacks like Java apps running in IBM websphere. There are…
I agree about the shepherding role. But, I'm not sure about how hands-off the shepherding will be at well-run companies. I work on pricing software for a regulated industry. If I were to start over again from scratch on…
looks like it worked :)
“ Name another service that does what they do for FREE (and no ads) at this scale” and is reliable is the question
Hopefully because they’re focusing on having relevance in the future. It’s sad as a gen Xer to see Japanese electronics companies fading away.
Same here. Same with not downing the harshest rye whiskey again. Executive function something or other …
It isn’t racist to acknowledge that some schools select students based on factors other than demonstrated aptitude. In the last ten years it was a fad in the US to not require SAT scores until folks realized this was…
Maybe I’m missing something. How is codex desktop more feature rich that Claude desktop? Aren’t they about the same now? I actually found codex desktop cloud mode to be more difficult to use in the codex desktop vs…
Ive been using 5.6 sol xhigh everyday since it launched. It does feel slower than it used to be. Fast mode seems useless now - burns through tokens without a noticeable speed gain for me now.
"So we are going back to blue books..." exactly. And, it is also efficient. You just need a proctor to ID the test taker and make sure they aren't using AI to cheat in order to verify that the student learned something…
charge more for it. its fine if free customers dont have service. dont break paying customers
gitlab self hosted though? no outage. i remember equivalent functionality at my last job taking this path
curious about methodology. ive seen them post results for claude/codex when they only ran over benchmarks 3 times per model...
how hard could it be these days for a mid to large size eng company to have their own gitlab/hub type solution hosted in aws
#0-nines
are you sure its actually two nines and not zero?
WTF. Outage every month if not multiple. Even for paying enterprise plans. Also, doesn't even have RAG offering.
Well then, the credible, perv glasses wearing authority on AI has spoken
Re the part about the brain associating how you feel at the moment / how you think you are performing with how you judge yourself as a person: I recommend reading Steve Hayes work / ACT therapy including this video for…
Sounds like you need to switch jobs.