ncnvrgnt
@inconvergent.net
procedural & graphics artist
parallel programmer
tired
website | https://inconvergent.net
gallery | https://img.inconvergent.net/plot/
a game | https://boxtype.app
store | https://buy.inconvergent.net
parallel programmer
tired
website | https://inconvergent.net
gallery | https://img.inconvergent.net/plot/
a game | https://boxtype.app
store | https://buy.inconvergent.net
Pinned
ncnvrgnt
@inconvergent.net
· Jun 6
i published a first version of my browser puzzle/platform game (+ level editor). you can play it for free (w/a keyboard) at boxtype.app. levels are encoded in the url; if you make something you like you can share it with anyone
i'd love to hear what you think!
i'd love to hear what you think!
Reposted by ncnvrgnt
very helpful debug view is helpful
January 6, 2026 at 1:48 PM
very helpful debug view is helpful
the only thing (generative) AI has "democratised" is CEO brain
January 6, 2026 at 5:45 PM
the only thing (generative) AI has "democratised" is CEO brain
MCP seems like a declaration of failure of the main product. if the product worked, why do you need MCP?
January 6, 2026 at 5:08 PM
MCP seems like a declaration of failure of the main product. if the product worked, why do you need MCP?
very helpful debug view is helpful
January 6, 2026 at 1:48 PM
very helpful debug view is helpful
Reposted by ncnvrgnt
This “multitool promise” phrasing is an excellent one, and the thread a good explainer of what I’m on about every time I go “IT DOES NOT UNDERSTAND SEMANTICS”
(Which makes it baffling that coders fall for this shit. We’re supposed to understand semantics, not do educated surface-level guesswork)
(Which makes it baffling that coders fall for this shit. We’re supposed to understand semantics, not do educated surface-level guesswork)
Let me provide some context here: GenAI shills have over promised GenAI and their cousin platforms as the digital equivalent of a multitool that can be applied reliably to solve almost any problem encountered in the world. Moreover, they've convinced the public that this is the case.
"GenAI is the futu--
January 6, 2026 at 12:05 PM
This “multitool promise” phrasing is an excellent one, and the thread a good explainer of what I’m on about every time I go “IT DOES NOT UNDERSTAND SEMANTICS”
(Which makes it baffling that coders fall for this shit. We’re supposed to understand semantics, not do educated surface-level guesswork)
(Which makes it baffling that coders fall for this shit. We’re supposed to understand semantics, not do educated surface-level guesswork)
one of the more desirable features of machine learning (models) is the ability to know (at least) something about why you were given a particular result/prediction. the generative AI being pushed on us at the moment does not have this ability at all
January 6, 2026 at 11:23 AM
one of the more desirable features of machine learning (models) is the ability to know (at least) something about why you were given a particular result/prediction. the generative AI being pushed on us at the moment does not have this ability at all
it is impossible to know if crushing orphans is good or bad
January 6, 2026 at 11:03 AM
it is impossible to know if crushing orphans is good or bad
Reposted by ncnvrgnt
the shadows are unusual today
January 5, 2026 at 10:57 AM
the shadows are unusual today
"i generated [many] lines of code, and much more complexity than i could have done on my own" isn't the selling point you seem to think it is
January 5, 2026 at 5:47 PM
"i generated [many] lines of code, and much more complexity than i could have done on my own" isn't the selling point you seem to think it is
Reposted by ncnvrgnt
LLM users are listing generated lines of code as arguments of quality (???) in much the same way that certain newsletter writers list their word counts ...
January 5, 2026 at 11:45 AM
LLM users are listing generated lines of code as arguments of quality (???) in much the same way that certain newsletter writers list their word counts ...
Reposted by ncnvrgnt
i suspect people severely underestimate the negative cognitive effects of using AI tools. even for programming tasks. and even if you are someone who gets real efficiency gains from those tools; perhaps especially then
November 28, 2025 at 7:07 PM
i suspect people severely underestimate the negative cognitive effects of using AI tools. even for programming tasks. and even if you are someone who gets real efficiency gains from those tools; perhaps especially then
LLM users are listing generated lines of code as arguments of quality (???) in much the same way that certain newsletter writers list their word counts ...
January 5, 2026 at 11:45 AM
LLM users are listing generated lines of code as arguments of quality (???) in much the same way that certain newsletter writers list their word counts ...
too bad we won't see the research on what makes people predisposed to LLM psychosis and/or addiction in a while. we need more subjects, after all
January 5, 2026 at 11:27 AM
too bad we won't see the research on what makes people predisposed to LLM psychosis and/or addiction in a while. we need more subjects, after all
the shadows are unusual today
January 5, 2026 at 10:57 AM
the shadows are unusual today
nothing communicates that i can be trusted with sophisticated computing equipment quite like the fact that i spend it making this
January 4, 2026 at 10:44 PM
nothing communicates that i can be trusted with sophisticated computing equipment quite like the fact that i spend it making this
showing some SVG paths that have been moved around inside inkscape. the paths are drawn from the mesh with a greedy path planner
January 4, 2026 at 9:03 PM
showing some SVG paths that have been moved around inside inkscape. the paths are drawn from the mesh with a greedy path planner
having Linux on your computer is the moderate position
January 4, 2026 at 6:00 PM
having Linux on your computer is the moderate position
we might have a problem.
or the camera could be upside down.
how can you even know at this point.
or the camera could be upside down.
how can you even know at this point.
January 4, 2026 at 1:29 PM
we might have a problem.
or the camera could be upside down.
how can you even know at this point.
or the camera could be upside down.
how can you even know at this point.
Reposted by ncnvrgnt
I've had to revise the Black Ice timeline a few times because events I had listed in the 30s started happening in the 20s
January 4, 2026 at 3:47 AM
I've had to revise the Black Ice timeline a few times because events I had listed in the 30s started happening in the 20s
william gibson must be so tired of revising his next novel because reality catches up with him
January 3, 2026 at 10:02 PM
william gibson must be so tired of revising his next novel because reality catches up with him
Reposted by ncnvrgnt
many current problems began when people started trusting computer. you must never trust computer. computer is the machine that hates you
January 1, 2026 at 10:13 AM
many current problems began when people started trusting computer. you must never trust computer. computer is the machine that hates you
so we are not aiming for any improvement over last year, then
January 3, 2026 at 11:00 AM
so we are not aiming for any improvement over last year, then
