For you as supervisor not much has changed. Before you had juniors not having full context, hesitating to ask, focused on code. Now lying agents, pretending to have phd grade but taking dumb choices here and there. Trust but verify as Lenin was saying.
Engineering management after the cost of code collapsed
201–210 of 217 posts
Re: Engineering management after the cost of code collapsed
#202The cost of code actually increased; code debt is being accumulated faster than we can clean it up.
Business never cared about code debt, LLM helps building features faster into production and from their POV, that's all that matters. If your engineers are "wasting time" optimizing artisan code whilst the competition has released their next version, they'll be told to use AI. The other reality is that by the time you figure out the right abstraction, business has already pivoted, or your feature will be rewritten ,…
I'm glad, and less surprised now, that most important libraries and infrastructure projects were done by volunteers.
Imagine a world where those only get done for a business incentive by people with this hard-reality business mindset. Or not done at all, reinvented at every place, no transferable knowledge for developers.
Re: Engineering management after the cost of code collapsed
#203Earlier quoted context omitted.
This is so accurate. I am a fairly senior individual contributor and its been more than a year since I saw a good design doc or quality design discussion. Before you can even question the design, someone has generated a 50k line prototype and already made up their mind because of all the “you’re absolutely right..” and “X is exactly what you need..” from AI.
"Talk is cheap, show me the code" A working prototype beats endless discussions over paper
A working prototype like that is good if the debate is “is this possible?”, the problem above is more for problems where it’s possible and we’re arguing which path to take
Re: Engineering management after the cost of code collapsed
#204Earlier quoted context omitted.
Did you read the paper you are linking to? It records a 0% false positive rate for evaluation of human-authored controls, and fewer than 5% of the hybrid and humanized papers had their AI levels overestimated by pangram. It is not completely clear from the data, but it seems that depending on whether the n=2 overestimates were “100% ai generated” assessments, the study you link says pangram’s 100% AI assessments were…
My citation was correct. To question to ask yourself: For my use case, is it okay that Pangram can't reliably tell the difference between "100% AI generated" and "AI assisted"?
Your citation was incorrect. The data also shows that pangram reliably tells the difference between those two cases. The statistic you made up or hallucinated is not supported by your link, and you’re being disingenuous here.
Re: Engineering management after the cost of code collapsed
#205I think most of this is correct, in spite of potentially being built on a bad assumption. The assumption is that LLMs should be writing the code and human engineers reviewing and verifying the LLM output. And that this pushes the cost of producing down. And I fundamentally disagree with that. Every time I ask LLMs to write code, even with Opus 4.8 (haven't tried it with Opus 5 yet), what I get ends up being totally r…
There is the flip side of where the code is no longer read by humans and is becoming the prevailing way software is shipped in tiny businesses. You don't need to code to be maintainable since you will never maintain it, the AI will and the quality will naturally improve as models improve. For example 5.6 sol and fable are showing signs where you can feed garbage in and it will spit out something pretty decent, defini…
I find this to work for a V1 or product with only a few users but not sure this will scale. At least you still need software engineers reviewing the functionality and giving the AI hints at what to do.
Re: Engineering management after the cost of code collapsed
#206Earlier quoted context omitted.
Hard to say what caused what, but the internet seems to mistake verbosity for authority, and so does AI.
A tech comm course I took in college was graded on two 20-page papers and accompanying 5-minute presentations. That was like 10k written words in a single semester. It was a challenging class, and gave me substantial sense of accomplishment, just to hand in completed work. Similar to a functioning side project in the 5-10k LOC range. Announcing something that worked a year ago, was laudable, even if not profitable. I…
> I don't think that's widely recognized yet.
The recognition is represented by one word: Slop.
Re: Engineering management after the cost of code collapsed
#207Pangram reports this post was 100% AI generated.
Pangram's marketing always reminds me of Anchorman's Sex Panther cologne: "They've done studies, you know. Sixty percent of the time, it works every time." Pangram's "100% AI generated" claims are right 65% of the time. https://link.springer.com/article/10.1007/s40979-026-00226-w
As opposed to Colt 45, which works every time.
Re: Engineering management after the cost of code collapsed
#208Earlier quoted context omitted.
My citation was correct. To question to ask yourself: For my use case, is it okay that Pangram can't reliably tell the difference between "100% AI generated" and "AI assisted"?
Your citation was incorrect. The data also shows that pangram reliably tells the difference between those two cases. The statistic you made up or hallucinated is not supported by your link, and you’re being disingenuous here.
The result I'm citing is very clearly presented in both a table and in supporting copy. It's even highlighted in the summary. Search for "65".
Re: Engineering management after the cost of code collapsed
#209Earlier quoted context omitted.
Your citation was incorrect. The data also shows that pangram reliably tells the difference between those two cases. The statistic you made up or hallucinated is not supported by your link, and you’re being disingenuous here.
The result I'm citing is very clearly presented in both a table and in supporting copy. It's even highlighted in the summary. Search for "65".
Thank you for pointing that out. But it still doesn’t support what you said, which is that only 65% of pangram’s fully AI claims are accurate. The “65% strict accuracy” rating is followed in the same sentence by a “97.5% inclusive accuracy” rating for the same reasons I pointed out above: pangram is systematically underrating the percentage AI in content, and it is tuned against producing false positives. The 32.5% in between the inclusive and strict accuracy ratings is entirely due to it estimating that 12 fully ai generated works were 60-80% AI, not the reverse. You can see that in table 4.
Re: Engineering management after the cost of code collapsed
#210Earlier quoted context omitted.
"Talk is cheap, show me the code" A working prototype beats endless discussions over paper
Talk is cheap, that's why you do it before sinking months of man-effort into building a turd.
This. Exactly this. Most ideas are bad ideas. Now noone can stop all of them from being sent to prod for everyone to deal with.