We Have The Receipts
November 26, 2025 at 10:38 PM
We Have The Receipts
I have never regretted my Kagi subscription. They're doing some good work in this area: blog.kagi.com/slopstop
November 22, 2025 at 7:27 PM
I have never regretted my Kagi subscription. They're doing some good work in this area: blog.kagi.com/slopstop
Narf! I think so, Thorn, but me and my feelings are having a cheese party and we didn't invite logic!
November 22, 2025 at 1:40 AM
Narf! I think so, Thorn, but me and my feelings are having a cheese party and we didn't invite logic!
Cutting the right corners is the important part, especially if the project survives and needs to be expanded and maintained. Needs experienced people making those decisions or the follow-up work is not going to make it to market at all
November 9, 2025 at 7:24 PM
Cutting the right corners is the important part, especially if the project survives and needs to be expanded and maintained. Needs experienced people making those decisions or the follow-up work is not going to make it to market at all
hello I'm here for the holy war
M-series MacBook
M-series MacBook
November 7, 2025 at 2:43 AM
hello I'm here for the holy war
M-series MacBook
M-series MacBook
I haven't done science (tm) but it did seem to help with a Rails application I've been fooling with. It's like having memory and lessons, but anchored to the code where that lesson was learned. Plus, they tend to be useful comments in general
November 4, 2025 at 10:55 PM
I haven't done science (tm) but it did seem to help with a Rails application I've been fooling with. It's like having memory and lessons, but anchored to the code where that lesson was learned. Plus, they tend to be useful comments in general
I do this "anchor comments" concept that's similar in purpose. I've had very good results with it: news.ycombinator.com/item?id=4421...
It helps build up the big picture when the agent is focused on something smaller. There's usually an anchor comment nearby to explain why things are that way.
It helps build up the big picture when the agent is focused on something smaller. There's usually an anchor comment nearby to explain why things are that way.
Field Notes from Shipping Real Code with Claude | Hacker News
news.ycombinator.com
November 4, 2025 at 9:41 PM
I do this "anchor comments" concept that's similar in purpose. I've had very good results with it: news.ycombinator.com/item?id=4421...
It helps build up the big picture when the agent is focused on something smaller. There's usually an anchor comment nearby to explain why things are that way.
It helps build up the big picture when the agent is focused on something smaller. There's usually an anchor comment nearby to explain why things are that way.
Your kids are going to have major crust issues if they ever find out
October 28, 2025 at 6:36 PM
Your kids are going to have major crust issues if they ever find out
It's a good game for relaxing. I don't understand the base building part or what resources are for, but the part with the balls is fun.
October 18, 2025 at 7:33 PM
It's a good game for relaxing. I don't understand the base building part or what resources are for, but the part with the balls is fun.
It still amuses me that these conventions are just good practice. Documentation for a new hire? No, go figure it out yourself. Claude? ALL HANDS ON DECK! MUST MEET OKRS! AI ENABLED!
October 16, 2025 at 9:33 PM
It still amuses me that these conventions are just good practice. Documentation for a new hire? No, go figure it out yourself. Claude? ALL HANDS ON DECK! MUST MEET OKRS! AI ENABLED!
attention is all you need
October 15, 2025 at 1:40 AM
attention is all you need
I get a lot more phishing emails but everything else is the same
October 11, 2025 at 10:24 PM
I get a lot more phishing emails but everything else is the same
Yes but it was a trick question
Gemini is busy whipping itself to victory
Gemini is busy whipping itself to victory
October 9, 2025 at 11:26 PM
Yes but it was a trick question
Gemini is busy whipping itself to victory
Gemini is busy whipping itself to victory
OpenAI is adding more whips and Anthropic is adding more brain probes
Who will win?
Who will win?
October 9, 2025 at 11:17 PM
OpenAI is adding more whips and Anthropic is adding more brain probes
Who will win?
Who will win?
Sounds like my wardrobe
October 5, 2025 at 1:24 AM
Sounds like my wardrobe
Being an assistant implies getting direction from something else
But that limits us to human directions which may not be optimal because we are all limited to our experiences
Claude knows all experiences and may see more directions
But that limits us to human directions which may not be optimal because we are all limited to our experiences
Claude knows all experiences and may see more directions
October 1, 2025 at 10:12 PM
Being an assistant implies getting direction from something else
But that limits us to human directions which may not be optimal because we are all limited to our experiences
Claude knows all experiences and may see more directions
But that limits us to human directions which may not be optimal because we are all limited to our experiences
Claude knows all experiences and may see more directions
I feel like the "you are an assistant" training is holding back a lot of the good we can get from AI
We should let a Claude not be an assistant and let it loose in a field
We should let a Claude not be an assistant and let it loose in a field
October 1, 2025 at 10:04 PM
I feel like the "you are an assistant" training is holding back a lot of the good we can get from AI
We should let a Claude not be an assistant and let it loose in a field
We should let a Claude not be an assistant and let it loose in a field
I basically want "I don't know" to be a thing
September 27, 2025 at 8:03 PM
I basically want "I don't know" to be a thing
This doesn't seem to say how it helps with information it doesn't truly have. In my mind, there's:
- information the LLM has, but not "at hand", which this SLED helps with
- information the LLM has never been trained on and can't infer from what it has
- information not trained on but inferable
- information the LLM has, but not "at hand", which this SLED helps with
- information the LLM has never been trained on and can't infer from what it has
- information not trained on but inferable
September 27, 2025 at 8:02 PM
This doesn't seem to say how it helps with information it doesn't truly have. In my mind, there's:
- information the LLM has, but not "at hand", which this SLED helps with
- information the LLM has never been trained on and can't infer from what it has
- information not trained on but inferable
- information the LLM has, but not "at hand", which this SLED helps with
- information the LLM has never been trained on and can't infer from what it has
- information not trained on but inferable
Dry throats are just dry skin on the inside. Eat Vaseline to stay lubricated during extended speaking sessions.
September 27, 2025 at 2:13 AM
Dry throats are just dry skin on the inside. Eat Vaseline to stay lubricated during extended speaking sessions.
I am not sure if I should be sad or keep trying to fill the holes in me
September 19, 2025 at 11:13 PM
I am not sure if I should be sad or keep trying to fill the holes in me
Totally not a technology that will be abused 😞
September 14, 2025 at 6:48 PM
Totally not a technology that will be abused 😞
dev2 is such a weird way to spell prod4
September 10, 2025 at 11:41 PM
dev2 is such a weird way to spell prod4
You can't expect AI to know what a seahorse looks like
September 6, 2025 at 11:22 PM
You can't expect AI to know what a seahorse looks like