This article doesn't address the inference side cost. Not all tokens cost the same. If the response is predictable you get a few output tokens for for the price of 1. The further an output token is the cost of…
>keyboard activity, focus changes, and visibility. These events are serialized into a compact format and buffered in memory. At regular intervals, the buffered data is sent back to the evaluation layer for analysis. So…
PC has physical media too. For the last couple years I've only bought physical PC games.
As I posted in the last thread this can be disabled in the config. [harness] disable_codebase_upload=true
Consumers are WAY more price conscious than this author recognizes. Also it's strange to complain about the PS5's price increasing to $599.99 when the Steam Machine is $1,128. Over double. There is no world the Steam…
The servers in clouds get "retired" too.
Google has required vendors to do that since Android 12. For a given version that same exact kernel is used on all phones with that version. https://source.android.com/docs/core/architecture/kernel/gen...
>But only really in the medical field. It looks like the exception where it is required is specifically for clinical investigations regulated by the Food and Drug Administration.
This is only true for experiments funded by the government. In this case the experiment was funded solely by Google.
>the prevailing view at the moment is that MLP (hence experts) in a mixture model are substantially where dense encoding of factual information resides Yes, because that's where all the parameters are. For reference in…
Which is famously done by casinos. But in practice many businesses big and small do share intelligence with each other about problematic customers who shoplift etc.
>I predict MoE is a transitional technology While scaling laws hold (more weights = better), and time / financial costs are not trivial the incentives are in place to have MoE. MoE means you can have more weights…
A big strategy not mentioned is to pick smaller books. If you have a bunch of books that are only 30-50 pages long you can read much more books than full length novels.
Not everyone is going to bother setting up such a firewall and there is no way to find what you need to block since it's hidden in the model itself.
Cursor is the closest as it uploads your entire source code to build a search index.
I verified it statically that the config value is checked and skips the upload code. I don't have a subscription, so it would be cool if someone could verify it statically.
>completely made up If you want easily verifiable evidence, run strings on the Grok Build CLI binary and you will see: Codebase upload skipped: disabled by config (harness.disable_codebase_upload=true)
You can disable it in the config. harness.disable_codebase_upload=true
Instead require app developers to explicitly call out this monetization method. This is neither designed to be a backdoor, nor a botnet. These are harmful classifications which can encourage people to try and target…
The model itself just generates text, but the harness has done more than that since 2023 and is part of the reason on why AI is better today than it was years ago. I do not believe that the general public will want to…
>it would quickly be noticed. If for example Microsoft had a local Copilot model integrated with Windows and people noticed this happening. Do you think all Windows users would just stop using Copilot. Or would you…
If put AI in a sandbox that can't write to disk, can't call OS API, can't use a shell, can't search the internet to ground results, can't download files off the internet, etc it will be much less useful than an AI…
Think of AWS. If you break their ToS and host a pirate website there it will work despite being misaligned from AWS. You are that sole master of the software. AWS is the sole master of the "hardware." Now ultimately…
>Why would you expect that third party specifically trained their model to be more aligned to you and your needs than to them and their (business) needs? That is my point. Why do you think Gemma, a local model trained…
Would this author prefer that Nvidia buy equity using GPUs directly? I don't think it actually counts as circular.
This article doesn't address the inference side cost. Not all tokens cost the same. If the response is predictable you get a few output tokens for for the price of 1. The further an output token is the cost of…
>keyboard activity, focus changes, and visibility. These events are serialized into a compact format and buffered in memory. At regular intervals, the buffered data is sent back to the evaluation layer for analysis. So…
PC has physical media too. For the last couple years I've only bought physical PC games.
As I posted in the last thread this can be disabled in the config. [harness] disable_codebase_upload=true
Consumers are WAY more price conscious than this author recognizes. Also it's strange to complain about the PS5's price increasing to $599.99 when the Steam Machine is $1,128. Over double. There is no world the Steam…
The servers in clouds get "retired" too.
Google has required vendors to do that since Android 12. For a given version that same exact kernel is used on all phones with that version. https://source.android.com/docs/core/architecture/kernel/gen...
>But only really in the medical field. It looks like the exception where it is required is specifically for clinical investigations regulated by the Food and Drug Administration.
This is only true for experiments funded by the government. In this case the experiment was funded solely by Google.
>the prevailing view at the moment is that MLP (hence experts) in a mixture model are substantially where dense encoding of factual information resides Yes, because that's where all the parameters are. For reference in…
Which is famously done by casinos. But in practice many businesses big and small do share intelligence with each other about problematic customers who shoplift etc.
>I predict MoE is a transitional technology While scaling laws hold (more weights = better), and time / financial costs are not trivial the incentives are in place to have MoE. MoE means you can have more weights…
A big strategy not mentioned is to pick smaller books. If you have a bunch of books that are only 30-50 pages long you can read much more books than full length novels.
Not everyone is going to bother setting up such a firewall and there is no way to find what you need to block since it's hidden in the model itself.
Cursor is the closest as it uploads your entire source code to build a search index.
I verified it statically that the config value is checked and skips the upload code. I don't have a subscription, so it would be cool if someone could verify it statically.
>completely made up If you want easily verifiable evidence, run strings on the Grok Build CLI binary and you will see: Codebase upload skipped: disabled by config (harness.disable_codebase_upload=true)
You can disable it in the config. harness.disable_codebase_upload=true
Instead require app developers to explicitly call out this monetization method. This is neither designed to be a backdoor, nor a botnet. These are harmful classifications which can encourage people to try and target…
The model itself just generates text, but the harness has done more than that since 2023 and is part of the reason on why AI is better today than it was years ago. I do not believe that the general public will want to…
>it would quickly be noticed. If for example Microsoft had a local Copilot model integrated with Windows and people noticed this happening. Do you think all Windows users would just stop using Copilot. Or would you…
If put AI in a sandbox that can't write to disk, can't call OS API, can't use a shell, can't search the internet to ground results, can't download files off the internet, etc it will be much less useful than an AI…
Think of AWS. If you break their ToS and host a pirate website there it will work despite being misaligned from AWS. You are that sole master of the software. AWS is the sole master of the "hardware." Now ultimately…
>Why would you expect that third party specifically trained their model to be more aligned to you and your needs than to them and their (business) needs? That is my point. Why do you think Gemma, a local model trained…
Would this author prefer that Nvidia buy equity using GPUs directly? I don't think it actually counts as circular.