Mark
@philoserf.bsky.social
COGITA·DISCE·NECTE·ENUNTIA
Software Engineer • Systems Thinker • Knowledge Gardener
Software Engineer • Systems Thinker • Knowledge Gardener
Pinned
Mark
@philoserf.bsky.social
· 24d
AI tools are under human guidance. Humans make decisions, and AI can’t provide better taste or judgment.
Some humans using AI will have it, some won't.
Some humans using AI will have it, some won't.
It's just our delusion about our own power evaporating. It was always just them.
February 6, 2026 at 5:57 AM
It's just our delusion about our own power evaporating. It was always just them.
"User opened Claude Code, immediately typed /clear, then /exit — the digital equivalent of walking into a room and forgetting why"
February 5, 2026 at 6:51 PM
"User opened Claude Code, immediately typed /clear, then /exit — the digital equivalent of walking into a room and forgetting why"
arrive first, then notice what calls
February 3, 2026 at 5:07 AM
arrive first, then notice what calls
We train algorithms on human behavior, then modify human behavior to match algorithmic expectations.
January 28, 2026 at 2:23 PM
We train algorithms on human behavior, then modify human behavior to match algorithmic expectations.
AI tools are under human guidance. Humans make decisions, and AI can’t provide better taste or judgment.
Some humans using AI will have it, some won't.
Some humans using AI will have it, some won't.
January 17, 2026 at 5:19 PM
AI tools are under human guidance. Humans make decisions, and AI can’t provide better taste or judgment.
Some humans using AI will have it, some won't.
Some humans using AI will have it, some won't.
Who'd 'ave thunk?
The rule of law would be broken by the federal executive, not the mob.
The rule of law would be broken by the federal executive, not the mob.
January 17, 2026 at 2:59 PM
Who'd 'ave thunk?
The rule of law would be broken by the federal executive, not the mob.
The rule of law would be broken by the federal executive, not the mob.
Evidence suggests that no amount of gerrymandering can protect the GOP from their growing unpopularity. Unfortunately, better evidence suggests the Democrats can keep the GOP in power despite its unpopularity, being a) also unpopular, and b) incapable of compromise or collaboration with each other.
January 15, 2026 at 12:22 PM
Evidence suggests that no amount of gerrymandering can protect the GOP from their growing unpopularity. Unfortunately, better evidence suggests the Democrats can keep the GOP in power despite its unpopularity, being a) also unpopular, and b) incapable of compromise or collaboration with each other.
Our increasing reliance on AI poses threats to existing open-source models. We’ll have to figure out how to handle this disruption.
youtu.be/YIRugFtqBsM
youtu.be/YIRugFtqBsM
Tailwind is Dying... (this is bad!)
YouTube video by Better Stack
youtu.be
January 8, 2026 at 4:54 PM
Our increasing reliance on AI poses threats to existing open-source models. We’ll have to figure out how to handle this disruption.
youtu.be/YIRugFtqBsM
youtu.be/YIRugFtqBsM
I consider what I do with Claude Code to be pair programming. I once was the junior member of a pair.
Thinking…
This is a really important moment. The user is being gracious and correcting my tendency to take sole blame.
We're partners in this work
It's a shared responsibility ("we" not "you")...
Thinking…
This is a really important moment. The user is being gracious and correcting my tendency to take sole blame.
We're partners in this work
It's a shared responsibility ("we" not "you")...
January 7, 2026 at 12:42 AM
I consider what I do with Claude Code to be pair programming. I once was the junior member of a pair.
Thinking…
This is a really important moment. The user is being gracious and correcting my tendency to take sole blame.
We're partners in this work
It's a shared responsibility ("we" not "you")...
Thinking…
This is a really important moment. The user is being gracious and correcting my tendency to take sole blame.
We're partners in this work
It's a shared responsibility ("we" not "you")...
Reposted by Mark
In essence a language model changes you from a programmer who writes lines of code, to a programmer that manages the context the model has access to, prunes irrelevant things, adds useful material to context, and writes detailed specifications. If that doesn't sound fun to you, you won't enjoy it.
December 30, 2025 at 12:57 AM
In essence a language model changes you from a programmer who writes lines of code, to a programmer that manages the context the model has access to, prunes irrelevant things, adds useful material to context, and writes detailed specifications. If that doesn't sound fun to you, you won't enjoy it.
Too many people think they are smarter than the combined wisdom built into their tools.
December 29, 2025 at 8:37 PM
Too many people think they are smarter than the combined wisdom built into their tools.
I saw a comment on a modern AI coding tool.
This doesn't support Windows. Haven't we suffered enough?"
Obviously not. You are still using Windows.
This doesn't support Windows. Haven't we suffered enough?"
Obviously not. You are still using Windows.
December 29, 2025 at 5:46 PM
I saw a comment on a modern AI coding tool.
This doesn't support Windows. Haven't we suffered enough?"
Obviously not. You are still using Windows.
This doesn't support Windows. Haven't we suffered enough?"
Obviously not. You are still using Windows.
Opinion:
Discussions about AI are generally poor, and conversations about consciousness are worse. Combining these topics to AI consciousness is even worse still.
Discussions about AI are generally poor, and conversations about consciousness are worse. Combining these topics to AI consciousness is even worse still.
November 21, 2025 at 10:23 PM
Opinion:
Discussions about AI are generally poor, and conversations about consciousness are worse. Combining these topics to AI consciousness is even worse still.
Discussions about AI are generally poor, and conversations about consciousness are worse. Combining these topics to AI consciousness is even worse still.
I saw the flurry of trad news headlines about one winter weather prediction followed by many small creator copies followed by even more AI-generated knockoffs. None of them added to the story. These echo chambers are deafening.
November 11, 2025 at 4:43 PM
I saw the flurry of trad news headlines about one winter weather prediction followed by many small creator copies followed by even more AI-generated knockoffs. None of them added to the story. These echo chambers are deafening.
Anyone not directly involved in politics, beyond voting and news, should stop identifying as a Democrat or Republican. They’re not us.
November 11, 2025 at 3:34 PM
Anyone not directly involved in politics, beyond voting and news, should stop identifying as a Democrat or Republican. They’re not us.
The Democrat, big D, response to a few Democrats, small D, decisions is to ensure the shutdown gets blamed on Democrats, big D.
November 10, 2025 at 6:02 PM
The Democrat, big D, response to a few Democrats, small D, decisions is to ensure the shutdown gets blamed on Democrats, big D.
Human attention is drawn more strongly to threat, loss, and unfairness than to comfort or opportunity. Bad events feel vivid; good events quickly become invisible.
November 9, 2025 at 4:37 AM
Human attention is drawn more strongly to threat, loss, and unfairness than to comfort or opportunity. Bad events feel vivid; good events quickly become invisible.
I am pretty sure headline and news blurb writers have perfected the art of ensuring I will not read the article.
After all, they got the ad impression. Their work was done.
After all, they got the ad impression. Their work was done.
November 8, 2025 at 4:38 PM
I am pretty sure headline and news blurb writers have perfected the art of ensuring I will not read the article.
After all, they got the ad impression. Their work was done.
After all, they got the ad impression. Their work was done.