Rubio has some big shoes to fill.
Oh, wait.
Oh, wait.
August 1, 2026 at 10:38 PM
Rubio has some big shoes to fill.
Oh, wait.
Oh, wait.
Jesus: woke corpo HR
June 18, 2026 at 10:13 PM
Jesus: woke corpo HR
Early 80s USA we got various foods (peanut butter, cheese, others I don't remember) through a government program. I remember it being called DPA or similar. It came in plain white containers with black lettering similar to what's depicted here.
June 18, 2026 at 12:33 AM
Early 80s USA we got various foods (peanut butter, cheese, others I don't remember) through a government program. I remember it being called DPA or similar. It came in plain white containers with black lettering similar to what's depicted here.
History shows again and again how nature points out the folly of man.
Well, man babies in this case.
Well, man babies in this case.
June 16, 2026 at 2:30 PM
History shows again and again how nature points out the folly of man.
Well, man babies in this case.
Well, man babies in this case.
Mister Rogers does not deserve an association with Rubio.
June 7, 2026 at 6:23 AM
Mister Rogers does not deserve an association with Rubio.
At least in my area (dystopian rural NC) my groceries are delivered by "contractors" from the same companies that deliver the burritos. The grocery store (Walmart) has my order for a week but the delivery itself is scheduled less than an hour before it's delivered.
June 2, 2026 at 7:57 AM
At least in my area (dystopian rural NC) my groceries are delivered by "contractors" from the same companies that deliver the burritos. The grocery store (Walmart) has my order for a week but the delivery itself is scheduled less than an hour before it's delivered.
As a small counterpoint, I am disabled, on SSI, and very much enjoy being able to get groceries delivered to me at a price I can pay.
June 2, 2026 at 6:38 AM
As a small counterpoint, I am disabled, on SSI, and very much enjoy being able to get groceries delivered to me at a price I can pay.
Have an angry upvote. I think all of the actors on the show did excellent jobs with what they were given but Dinklage was first among equals, as they say.
May 26, 2026 at 5:09 PM
Have an angry upvote. I think all of the actors on the show did excellent jobs with what they were given but Dinklage was first among equals, as they say.
are serviceable for some other tasks. The issue is non of those tasks are going to make infinite money.
It's like after WWII when companies tried to put jet engines into cars and they sucked. Except no one was stupid enough to tell people to break up with someone using a jet engine.
It's like after WWII when companies tried to put jet engines into cars and they sucked. Except no one was stupid enough to tell people to break up with someone using a jet engine.
May 25, 2026 at 4:32 PM
are serviceable for some other tasks. The issue is non of those tasks are going to make infinite money.
It's like after WWII when companies tried to put jet engines into cars and they sucked. Except no one was stupid enough to tell people to break up with someone using a jet engine.
It's like after WWII when companies tried to put jet engines into cars and they sucked. Except no one was stupid enough to tell people to break up with someone using a jet engine.
LLM's are really good at being coding assistants and sorting through data when paired with a human who knows what they're doing.
They are better than alternatives for any natural language translation that can't be put in front of a human translator.
And there are a few other areas where they
They are better than alternatives for any natural language translation that can't be put in front of a human translator.
And there are a few other areas where they
May 25, 2026 at 4:30 PM
LLM's are really good at being coding assistants and sorting through data when paired with a human who knows what they're doing.
They are better than alternatives for any natural language translation that can't be put in front of a human translator.
And there are a few other areas where they
They are better than alternatives for any natural language translation that can't be put in front of a human translator.
And there are a few other areas where they
Not complaining that you posted it. I just think these accounts want everyone to believe they're more important than they actually are.
May 23, 2026 at 7:37 PM
Not complaining that you posted it. I just think these accounts want everyone to believe they're more important than they actually are.
Do the people who do this matter? They're a vanishingly small portion of the electorate and not part of the "Stop Republicans from turning America into a white supremacist single party dictatorship" demographic.
I think all they're doing is poisoning the term "left."
I think all they're doing is poisoning the term "left."
May 23, 2026 at 7:31 PM
Do the people who do this matter? They're a vanishingly small portion of the electorate and not part of the "Stop Republicans from turning America into a white supremacist single party dictatorship" demographic.
I think all they're doing is poisoning the term "left."
I think all they're doing is poisoning the term "left."
The models I'm trialing now are Google's Gemma4, Qwen3.6, and a community finetune of an older Qwen model that is supposed to be as competent as the new Qwen models despite being a much smaller model. I think things like that can (possibly) survive an AI winter even if large studios never use them.
May 18, 2026 at 10:09 PM
The models I'm trialing now are Google's Gemma4, Qwen3.6, and a community finetune of an older Qwen model that is supposed to be as competent as the new Qwen models despite being a much smaller model. I think things like that can (possibly) survive an AI winter even if large studios never use them.
My non-expert impression is that some of the improvement in small models is due to newer, better models, but the tooling around them (standardized tool calls, systems that help work around context issues like forgetting and ignoring instructions) has advanced quickly.
May 18, 2026 at 10:06 PM
My non-expert impression is that some of the improvement in small models is due to newer, better models, but the tooling around them (standardized tool calls, systems that help work around context issues like forgetting and ignoring instructions) has advanced quickly.
I am not adverse to the tech bros getting what they deserve and I would expect that formal languages will dominate at most levels, I think that less formal approaches will continue to exist.
May 18, 2026 at 10:00 PM
I am not adverse to the tech bros getting what they deserve and I would expect that formal languages will dominate at most levels, I think that less formal approaches will continue to exist.
Yeah, I'm not wasting any more time on you.
May 18, 2026 at 8:07 PM
Yeah, I'm not wasting any more time on you.
40 years ago, writing assembly code was a vital skill but compilers got so good at that it no longer is. 40 years from now there are still going to be people who are software engineers and programmers, even if the way they interact with software ends up looking quite different than it does today.
May 18, 2026 at 8:01 PM
40 years ago, writing assembly code was a vital skill but compilers got so good at that it no longer is. 40 years from now there are still going to be people who are software engineers and programmers, even if the way they interact with software ends up looking quite different than it does today.
LLM's definitely have some huge ethical hurdles so I'm not even going to disagree with that. It's simply irrelevant to what you actually claimed. I wouldn't have replied at all had you just stated that.
May 18, 2026 at 5:48 PM
LLM's definitely have some huge ethical hurdles so I'm not even going to disagree with that. It's simply irrelevant to what you actually claimed. I wouldn't have replied at all had you just stated that.
Look, the original point you made is wrong. Your 1) point is also wrong. LLM's do that, not always well and I'd argue never as good as really good humans. And 2) is probably what you should have lead with, if that's what you believe instead of making (easily provable) false statements.
May 18, 2026 at 5:46 PM
Look, the original point you made is wrong. Your 1) point is also wrong. LLM's do that, not always well and I'd argue never as good as really good humans. And 2) is probably what you should have lead with, if that's what you believe instead of making (easily provable) false statements.
than telling the computer to build you an app.
Programming has more in common with brick laying than it does with creative writing which is probably why LLM's are better there.
Programming has more in common with brick laying than it does with creative writing which is probably why LLM's are better there.
May 18, 2026 at 4:26 PM
than telling the computer to build you an app.
Programming has more in common with brick laying than it does with creative writing which is probably why LLM's are better there.
Programming has more in common with brick laying than it does with creative writing which is probably why LLM's are better there.
"Vibe coding" is iffy, but doable on something like Claude. Bouncing questions about code off the local models, or using Claude to help architect a project, generally feels pretty good to me as a person who used to write code. It feels like using a higher level language more
May 18, 2026 at 4:25 PM
"Vibe coding" is iffy, but doable on something like Claude. Bouncing questions about code off the local models, or using Claude to help architect a project, generally feels pretty good to me as a person who used to write code. It feels like using a higher level language more
I think this has changed quite a bit in the past few months. I'm using local models and they're quite proficient at writing functions and even entire files so long as you're able to act as a project manager. I've taken to referring to the models I use as pseudo-code interpreters.
May 18, 2026 at 4:21 PM
I think this has changed quite a bit in the past few months. I'm using local models and they're quite proficient at writing functions and even entire files so long as you're able to act as a project manager. I've taken to referring to the models I use as pseudo-code interpreters.
I use a local llm for coding, along with small amounts of Claude and Gemini. I really wish everyone would stop trying to turn my pseudo-code interpreter into a real boy.
May 15, 2026 at 4:41 PM
I use a local llm for coding, along with small amounts of Claude and Gemini. I really wish everyone would stop trying to turn my pseudo-code interpreter into a real boy.
And I think that's what the article is saying, maybe?
Sorry, I tried to be clear and failed.
Sorry, I tried to be clear and failed.
May 13, 2026 at 10:05 PM
And I think that's what the article is saying, maybe?
Sorry, I tried to be clear and failed.
Sorry, I tried to be clear and failed.