ghost of john gode @ghostofjohngode.bsky.social · 16/10/20262 weeks or less until they announce "the unknown" as the new white house press secretary 062 6
ghost of john gode @ghostofjohngode.bsky.social · 09/09/2026wikipedia was never slop, but you couldn't cite it the same way as you would cite peer-reviewed research or published journalism and you still can't! idiot guy 010 1
ghost of john gode @ghostofjohngode.bsky.social · 05/08/2026on top the free advertising by (probably!) accident, I love that it's making it actively harder to find old posts, too 010 1
ghost of john gode @ghostofjohngode.bsky.social · 21/05/2026worked at a place where the no-shit answer to this problem was to just have the LLM tools do the review/debugging/etc and it broke the shit out of everything lol (source: third-bit.com/2026/05/20/t...) 210 2
ghost of john gode @ghostofjohngode.bsky.social · 11/05/2026I don't think you actually understood the posts here, but I guess I'm not surprised 100 1
ghost of john gode @ghostofjohngode.bsky.social · 30/03/2026walking out of Plato's cave, having a quick look around, and then deciding actually I'd rather go back in 000 0
ghost of john gode @ghostofjohngode.bsky.social · 12/02/2026shouldn't have left this guy at home, Guerin 000 0
ghost of john gode @ghostofjohngode.bsky.social · 09/02/2026these guys just can't help themselves but to crib from the 14 words jfc 000 0
ghost of john gode @ghostofjohngode.bsky.social · 08/02/2026this kind of thing should rightfully get you fired from being the dog catcher let alone managing the kind of wealth this guy deals with 100 1
ghost of john gode @ghostofjohngode.bsky.social · 07/02/2026@yeoldebuttpirate.bsky.social dawg I'm dying at this 000 0
ghost of john gode @ghostofjohngode.bsky.social · 01/02/2026of all the guys to be throwing around "creepy loser" accusations 010 1
ghost of john gode @ghostofjohngode.bsky.social · 31/01/2026here are the CEO and COO of my apartment 151 5
ghost of john gode @ghostofjohngode.bsky.social · 28/12/2025not the silliest but the most recent at least 0170 17
ghost of john gode @ghostofjohngode.bsky.social · 28/12/2025this industry, man, we're so cooked lol 000 0
ghost of john gode @ghostofjohngode.bsky.social · 18/12/2025ez way to tell this stuff is snake oil is that it gets presented with unqualified characterizations... trying to sell me on how innovative the AI functionality is in a settings interface is bozo shit 100 1
ghost of john gode @ghostofjohngode.bsky.social · 29/11/2025guess the ant got bored with autofellating? 000 0
ghost of john gode @ghostofjohngode.bsky.social · 29/10/2025I love everyone so I would never advocate for this but it's hard to not consider that people who think this would best serve the collective project of humanity by working to the death. obviously that's wrong because forced labor is bad, but you can't help but think about it 000 0
ghost of john gode @ghostofjohngode.bsky.social · 27/10/2025twice convicted sex offenders for andrew cuomo lol 000 0
ghost of john gode @ghostofjohngode.bsky.social · 27/10/2025moronic read of the last election cycle. the crypto guys bought the election with US dollars from VC and pyramid schemes, they didn't buy it with fucking ethereum 0252 25
ghost of john gode @ghostofjohngode.bsky.social · 23/10/2025learn faster maybe but certainly not deeper 020 2
ghost of john gode @ghostofjohngode.bsky.social · 23/10/2025did u know rick is almost translucent now 120 2
ghost of john gode @ghostofjohngode.bsky.social · 01/10/2025not gonna tolerate this shit on my timeline. if you're gonna post about baseball you gotta do it with the hater's mindset. save this theater kid squealing for your diary 120 2
ghost of john gode @ghostofjohngode.bsky.social · 11/09/2025really honoring the late pedophilia icon with this stat line, boys 000 0
![Measuring Only the Easy Half
LLMs make code generation faster, and that half is easy to measure. The other half is harder: time spent reviewing LLM-generated code for correctness, time lost debugging confidently wrong suggestions, security vulnerabilities introduced by plausible-looking but insecure code, and technical debt from suggestions that solved the immediate problem while ignoring the surrounding design. A study of GitHub Copilot’s code found that a substantial fraction of generated code contained security vulnerabilities, and that developers under time pressure accepted insecure suggestions at higher rates [Pearce2022]. A 2025 evaluation of five major LLMs found that none produced web application code meeting industry security standards [Dora2025]. A large-scale analysis of over 300,000 AI-authored commits found that more than 15% introduce at least one quality issue, and nearly a quarter of those issues persist in the codebase long-term [Liu2026]. Measuring only the inputs that go up while ignoring the costs that also rise is not measurement; it is marketing.](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:7bmyhqo3pwxtczomwpxia3px/bafkreia72tmwxip5q5mngosffk76wizzm7ul5jmoazklxqfsd27rbrosgm)
![csar 12 hours ago | prev | next [–]
Getting feedback on a plan or implementation is valuable because you get a fresh set of eyes. Using multiple models may help though it always feels a bit silly to me (if nothing else you’re increasing non-determinism because you know have to understand 2 LLM’s quirks).
But the “playing house” approach of experts is somewhere between pointless and actively harmful. It was all the rage in June and I thought people abandoned that later in the summer.
If you want the model to eg review code instead of fixing things, or document code without suggesting improvements (for writing docs), that’s useful. But there’s. I need for all these personas.
reply
bahaAbunojaim 9 hours ago | parent | next [–]
The way it works is that each agent think independently, discuss the solution and each agent opinion then one will synthesize a solution.
reply](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:7bmyhqo3pwxtczomwpxia3px/bafkreidfqfxu74pay56bp2jzqsznhq2vfomi5fn67xjz3pssa762z3laem)
