I know I’ve been burning a lot of resources in the last week, down to the system running out of tokens. Think deeply mode, with each step up to 10 minutes. June will be much the same. Definitely worth it, in my current use case, but just how much expensive iron-minutes have I’ve burned? I would sure like to know.
My company has let us get Enterprise Copilot and has been pushing us all to “use AI”. So, I now use it as a a semi-functional search for SharePoint/Outlook/Files on an almost daily basis. I also ask it questions about Microsoft documentation on a regular basis. I wonder how long it’s going to be until they yank my license.
Yeah my 0365 shoves it in the users faces, they ask it for some stuff until the honeymoon is over then all of a sudden it shits the bed when you try to make more than 2 images a day or plan a small project.
Think deeper in Copilot is apparently o3-mini, or it used to be until recently. GPT 5.5 (and some other versions) are also an option, so I should try using these next week.
If you have a high VRAM GPU (like 16GB or more) you can try running some model locally with LMStudio. You won’t get the same level as cloud models but depending on use case it might be good enough.
That’s poor but not horrible. LMStudio has Vulkan compute backend option that works on almost any card. You can also do partial GPU/RAM split giving you total of 72GB and increased performance vs CPU only.
Big companies license Copilot for less than 25 usd/month per seat. Don’t tell me it covers the ops cost, even for mixed calc.
Im June they are switching Copilot to metered usage . People are going to be out of credits on the third day.
I know I’ve been burning a lot of resources in the last week, down to the system running out of tokens. Think deeply mode, with each step up to 10 minutes. June will be much the same. Definitely worth it, in my current use case, but just how much expensive iron-minutes have I’ve burned? I would sure like to know.
My company has let us get Enterprise Copilot and has been pushing us all to “use AI”. So, I now use it as a a semi-functional search for SharePoint/Outlook/Files on an almost daily basis. I also ask it questions about Microsoft documentation on a regular basis. I wonder how long it’s going to be until they yank my license.
Apparently they’ll notice when you reach half a billion dollars
Only if they actually use it
Yeah my 0365 shoves it in the users faces, they ask it for some stuff until the honeymoon is over then all of a sudden it shits the bed when you try to make more than 2 images a day or plan a small project.
Depends how big the model is. Smaller models can be dirt cheap.
Think deeper in Copilot is apparently o3-mini, or it used to be until recently. GPT 5.5 (and some other versions) are also an option, so I should try using these next week.
If you have a high VRAM GPU (like 16GB or more) you can try running some model locally with LMStudio. You won’t get the same level as cloud models but depending on use case it might be good enough.
I only have an Radeon RX Vega 56 8GB HBM2, so limited to CPU-only. 64 GB memory there at least. Not upgrading unless hardware dies.
That’s poor but not horrible. LMStudio has Vulkan compute backend option that works on almost any card. You can also do partial GPU/RAM split giving you total of 72GB and increased performance vs CPU only.
Hmm, have to try it sometime when the hot phase at work is over. Thanks!