You're forced to document every check and thought related to development until you're wiring a manual on how to be laid off (unless self employed or otherwise critical). Which is fine. One should write quality docs and…
Russian diminutives, making nicknames much harder to track for those not familiar with the culture and language. Vladimir is Vova is Volodya, same person. Then other parts of their full name may have variations…
In my experience it's more the Russian diminutives across the language that cause issues. If you're familiar with the nickname variations then it's easy to follow. If you aren't you'll be taking notes and looking up why…
In my current use case I'm setting up a new Ubuntu server for hosting LLMs. I didn't take notes when setting it up last time around but want to document exactly what was required to pass on to coworkers trying something…
It sounds like they ship a fairly minimal and efficient set of tools or text editor. This would be unwanted bloat by those choosing them for those advantages.
Audit purposes for sure. How was this code/concept generated, what were the prompts/requirements, what thinking did the model complete, can this be replicated or repeated, etc. A vendor conference I was at a few weeks…
This seems like a perfect audit feature to flip back and train a model. Or ensure your human worker is working during business hours.
The best strategy is to figure out what you are going to do after smoking weed before you smoke weed. So, draft your prompt and send it before lighting up.
This was a primary goal (if not states) of USAID and related programs. Stem the causes of immigration, support stability, and create goodwill for the donor country. Still imperialistic and self serving in many ways, but…
This is in the direction of Mixture Of Export (MOE) setups. A trained 'router' sits on top of different expert models and routes work to the best/most efficient model for that task, and integrates the work into a whole…
I agree with this. The stack seems to become: LLM worker > Harness, Agents w/ skills > Human oversight/input This is similar in structure to many teams I work with, something like: Dev/SE/etc > PO > Manager/Director Or…
If any work is blocked/etc, refund all credits from that session/last X minutes. Minimum.
[dead]
US voters are defined by their apathy. Collective action is difficult even on areas most American's agree on due to shades of grey and implementation differences. It may be analogous to vaccine adoption, with…
Speed cameras, police with radar/etc, self reporting via insurance apps, controlled by GPS or speed limits in vehicles, etc The tech is there but not applied equally. Regulations could limit every car to 70mph, or the…
I want to try a hybrid setup of Gemma 4 E4B with lots of context for general, then Qwen 3.5 9B or larger for coding. Strix Halo set up this weekend, which may enable even larger Qwen models with tons of context.
If it's something like: - v4.5: 1x cost, 100% quality, 100% speed but maybe sometimes 80% speed because of load - v4.6: 3x cost, 105% quality, 80% speed most of the time depends - v4.7: 9x cost, 115% quality, 90% speed…
LM Studio Link is brilliant, outside their central login/auth requirement. Tailscale is the backbone, I think, so it makes sense but I'm sure a method with wireguard could exist and enable similar performance. the…
This plus the price different had me buy an AMD Strix Halo board last week. It seems the work with vLLM and training models could make the Spark worth the price difference, but before today's news I had the same thought…
If you can get desktop RTX 5070 performance, oodles of (v)RAM, and minimal power usage out of a thin and light mobile device it's a win. This is change. If you can afford it.
$20/day x 250 days per year x # devs/agents/etc = $$$. About $5k per dev at that daily use case. Enough to validate repurposing an existing workstation with enough RAM, or finding a used high VRAM GPU, or in my case…
On local models that cost power (post initial hardware cost), makes sense. My work is building this out and I think it's solid. But until we can use our own hardware and local models the long term cost is a big question…
A consumer "Standard GPU" could mean about a 6-8gb VRAM GPU still in support by the manufacturer, independent of CUDA/etc proprietary technology. Recent Steam hardware survey top GPU list is: - RTX 3060 (6 or 12gb VRAM)…
With the right harness/agent/skill setup and requirements work, this (any specificity) can become part of the workflow from early stages. Setting that up and making sure it works, at the early stages, is not what LLMs…
The Macbook Neo is about 500 nits but I'd bet the Asus laptop was more than $600 base. Looks like the Framework laptops depend on model/screen, between 4-500 but with the new 13 pro hitting 700 nits. For a user…
You're forced to document every check and thought related to development until you're wiring a manual on how to be laid off (unless self employed or otherwise critical). Which is fine. One should write quality docs and…
Russian diminutives, making nicknames much harder to track for those not familiar with the culture and language. Vladimir is Vova is Volodya, same person. Then other parts of their full name may have variations…
In my experience it's more the Russian diminutives across the language that cause issues. If you're familiar with the nickname variations then it's easy to follow. If you aren't you'll be taking notes and looking up why…
In my current use case I'm setting up a new Ubuntu server for hosting LLMs. I didn't take notes when setting it up last time around but want to document exactly what was required to pass on to coworkers trying something…
It sounds like they ship a fairly minimal and efficient set of tools or text editor. This would be unwanted bloat by those choosing them for those advantages.
Audit purposes for sure. How was this code/concept generated, what were the prompts/requirements, what thinking did the model complete, can this be replicated or repeated, etc. A vendor conference I was at a few weeks…
This seems like a perfect audit feature to flip back and train a model. Or ensure your human worker is working during business hours.
The best strategy is to figure out what you are going to do after smoking weed before you smoke weed. So, draft your prompt and send it before lighting up.
This was a primary goal (if not states) of USAID and related programs. Stem the causes of immigration, support stability, and create goodwill for the donor country. Still imperialistic and self serving in many ways, but…
This is in the direction of Mixture Of Export (MOE) setups. A trained 'router' sits on top of different expert models and routes work to the best/most efficient model for that task, and integrates the work into a whole…
I agree with this. The stack seems to become: LLM worker > Harness, Agents w/ skills > Human oversight/input This is similar in structure to many teams I work with, something like: Dev/SE/etc > PO > Manager/Director Or…
If any work is blocked/etc, refund all credits from that session/last X minutes. Minimum.
[dead]
US voters are defined by their apathy. Collective action is difficult even on areas most American's agree on due to shades of grey and implementation differences. It may be analogous to vaccine adoption, with…
Speed cameras, police with radar/etc, self reporting via insurance apps, controlled by GPS or speed limits in vehicles, etc The tech is there but not applied equally. Regulations could limit every car to 70mph, or the…
I want to try a hybrid setup of Gemma 4 E4B with lots of context for general, then Qwen 3.5 9B or larger for coding. Strix Halo set up this weekend, which may enable even larger Qwen models with tons of context.
If it's something like: - v4.5: 1x cost, 100% quality, 100% speed but maybe sometimes 80% speed because of load - v4.6: 3x cost, 105% quality, 80% speed most of the time depends - v4.7: 9x cost, 115% quality, 90% speed…
LM Studio Link is brilliant, outside their central login/auth requirement. Tailscale is the backbone, I think, so it makes sense but I'm sure a method with wireguard could exist and enable similar performance. the…
This plus the price different had me buy an AMD Strix Halo board last week. It seems the work with vLLM and training models could make the Spark worth the price difference, but before today's news I had the same thought…
If you can get desktop RTX 5070 performance, oodles of (v)RAM, and minimal power usage out of a thin and light mobile device it's a win. This is change. If you can afford it.
$20/day x 250 days per year x # devs/agents/etc = $$$. About $5k per dev at that daily use case. Enough to validate repurposing an existing workstation with enough RAM, or finding a used high VRAM GPU, or in my case…
On local models that cost power (post initial hardware cost), makes sense. My work is building this out and I think it's solid. But until we can use our own hardware and local models the long term cost is a big question…
A consumer "Standard GPU" could mean about a 6-8gb VRAM GPU still in support by the manufacturer, independent of CUDA/etc proprietary technology. Recent Steam hardware survey top GPU list is: - RTX 3060 (6 or 12gb VRAM)…
With the right harness/agent/skill setup and requirements work, this (any specificity) can become part of the workflow from early stages. Setting that up and making sure it works, at the early stages, is not what LLMs…
The Macbook Neo is about 500 nits but I'd bet the Asus laptop was more than $600 base. Looks like the Framework laptops depend on model/screen, between 4-500 but with the new 13 pro hitting 700 nits. For a user…