Replying to @DanBurmawy
Either this civilization ends Islam,
Or Islam will end this civilization.
Replying to @realGoodFriday @Branko1420
I mean, it all depends on whether you want to make your life more or less difficult....
15
Replying to @realGoodFriday @Branko1420
Or, hear me out on this, you can follow the bear TRACKS and see where he went...
103
Replying to @Gordosus12
Аааа блядь, напомнил про штабную крысу. Ну вот почему такая хуйня навсегда в мозгах отпечатывается?
161
Replying to @kav3rnus @VenomOnXx
Oh I thought they were literally beating the ghetto off him...
1,069
Replying to @designed_future @PermitZIPhvac
"Oh you're an interior designer? So you just like choose paint colors?" Uh, no. LOL.
1
5
Replying to @designed_future @PermitZIPhvac
and it just goes on and on.
Eventually, when each app gets to the level where I don't make any more adjustments, I have "yet another" app that makes Remix-Combine Candidates, and integrates them into 1 app that does 3-4 related things. But not just smooshing them together - actually reading the features and expected results, writing a CLEAN plan to make it, and creating a new one, as efficient as possible, but absorbing all the features and all the little details of the candidate apps.
IDK, it's my process, it works for me.
2
Replying to @designed_future @PermitZIPhvac
Example of 3 of such apps, each one fine-tuned for a very specific purpose (opened the log panel with details). Basically taking the approach of "do one thing but do it REALLY well", and TBH it's faster than waiting for the AI or the main app to respond. The Floorplan Preflight, for example, does a 4-story floorplan check in literally under 2 seconds - load, click, DONE.
1
1
9
"...In architecture, that means spotting the plausible-looking plan that violates something like egress, the claim about daylighting that ignores the site’s latitude or a precedent that doesn’t exist..." - absolutely correct, and it's the pitfall that claims too many over-eager practitioners.
Counterpoint to this: quantize and integrate the code references. You know what they are, your limit (as a human) is the attention to review 100s of possible applicability points. But that's EXACTLY what AI is good at - stupid repetitive things without getting bored or tired.
Case in point: one of the 50+ "helper apps" I've made for myself is a "Stair Designer", because I don't want to wait a week for an engineer to confirm code compliance. I'm not an engineer, I don't want to be an engineer. If I'm going to design a staircase for a custom home, I want to do it in a way that looks fantastic AND is buildable.
So I took the IBC and FBC texts and had Claude rip them apart into individual components and create a check routine that maps out each stair component to its relevant entry. So here's a staircase that's 100% compliant with the residential code, and if we had to reuse it in a commercial setting, there are only a couple of adaptations.
Why can't architects take the same approach? You HAVE the code references. CONNECT them to the workflow. Make them work IN CONTEXT. Have your system say IMMEDIATELY that something is wrong when you try to implement it.
People think "AI doesn't know how to make things compliant" and it's a problem.
It's not a problem. It's an opportunity.
Coming out of the recent elections, and bombarding the current news cycle, there is a lot of negative sentiment directed toward AI, much of it justified.
With so much negativity around AI, I wanted to focus exclusively on what AI is getting right.
I tallied the results - it turns out to be quite a lot.
More generally in medicine and science, then in architecture and the built environment, especially concerning early-stage energy, daylight and structural optimization, automated code-compliance checking, and fast documentation of at-risk heritage projects.
I also looked at why refusing to engage with AI doesn’t protect you from fears about jobs, the environment and cognitive offloading;
What people who have the most say over how AI is used in our field do differently;
Why architect Stephen Gallagher’s recent use of Opus 5.5 and now Fable 5.2 to produce professional-level permit-ready construction document sets ought to concern us and take seriously (I share my top 6 reasons);
And the strongest position we ought to take right now on AI use.
On this last one, current students, soon to be graduates and entry level designers have the most to gain - or lose - as the most pressing concern now is labor displacement at the entry level. Finding a job in architecture in the best of times is seldom easy and these are fraught times. What to do?
As a former firm principal and hiring manager, I've interviewed and hired hundreds of architects and recognize the moment when a job candidate is thrown off their game by an unexpected question. Don’t let an off-handed question about your AI use or your attitude about AI stump you or disrupt your momentum. To avoid this unwanted and unexpected outcome, I look at what candidates might expect to be asked about AI in the coming months.
Where we stand with AI and what AI gets (mostly) right theagenticarchitect.substack…
1
37
Replying to @PermitZIPhvac
Claude, but I've been training it to make plans and models for the last 6 months, including plenty of my own input + feeding it code books to digest and quantify, so it knows each *detail* and its application context. Also I have 1000s of rules by now, including 100s of layout-related ones like "labels must be on a cardinal direction, never on a random angle; if it's absolutely impossible to avoid a collision, then do a callout - a single intentional non-cardinal offset looks better than random placements". Which really comes into play on plans like these.
3
1
18
Replying to @Branko1420
Actually, "honey-knower", "ведать" is the ancient Russian for "to know", interestingly almost unchanged from Sanskrit (vedda). I.e. it's the animal who knows where the honey is. And if you carefully follow one, you can find the wild beehive, grab and domesticate it, and have honey for the village. Bears are useful :)
6
100
4,504
Replying to @alex_sasha2016 @vmarkov
От французского "beau monde", "красивый мир". Сборное название золотой молодёжи, олигархов, актёров, и т.д., которые тратят деньги пачками ради повыебнуться.
1
9
Replying to @PermitZIPhvac
Interesting how completely different methods converge on similar layouts. Maybe it's the data flow, though - 2 diagrams + notes + data tables tend to end up looking similar.
(Custom system I'm building: Fable 5.1 as orchestrator, Opus supervisors, Sonnet workers; 12K+ house rules, 1400 pages of ingested code specs; I swear it's getting better than the actual engineers we work with... (Still need them for the signature etc, but the work is done)).
1
1
28
Replying to @BullMooseMemes @kaimicahmills
Aren't you glad that you get to express your opinion on a fully organic, 100% natural social media network that grew on a developer tree?
17
Replying to @realMaalouf
We need bait squads. One guy pretends to be a target, the ferals come out and attack him, but what they don't expect is a dozen dudes coming around a corner and stress-testing baseball bats.
Do this enough times, the ferals will learn to fear.
1
1
165









