AI and the end of the road for Archicad

[gavinNZz]

AI tech is moving faster than anything we have seen before. The gains that can be made are so large there is huge money being poured into the development of AI in all industries. Architecture is no different. The video below shows where AI is already at so the next 5 yrs is going to see a big shift in the structure of Architectural practices around the world with a lot of labour intensive drafting processes being replaced with AI. Is GS ready for this transformation? Do they have a strategy in place to ensure they are still going to be relevant in an AI dominated world? Does Nemetschek actually care or will AC just be bleed dry until it becomes irrelevant?

AI will eventually produce the drawings and BIM models without the need for a piece of architectural software so the AC's and Revit's of this day in age will just fade away like the pen and paper before before them. 

While we complain about the functions, or lack thereof, in AC, I can see the reasoning for not throwing money at development when the return on investment won't have time to materialize as AI continues to disrupt our industry. Personally I can't see a future for AC, as it currently is, in this new world unless they embrace the AI wave and transform AC into an AI based piece of software. Given this type of transformation would run into the hundreds of millions I just can't see Nemetschek footing the bill.

How do other AC users see themselves placed in this new world? Are you going to be riding the AI wave or will you be left behind? It's a question we all need to be asking ourselves because AI is here and it will affect every aspect of our industry.

https://youtu.be/h-D5-6koRf8

[DGSketcher]

Some Qs looking beyond architecture:

1. Is AI capable of delivering the spark of inspiration for evolution, or does it risk analysing everything down to a simple binary decision?  8 billion copies is far more efficient than 8 billion unique solutions.

2. Is this a sustainable technology? If AI makes everyone's job redundant what do you see yourself doing? Is there a chance humanity will cease to exist as the jobless population is redundant to capitalism and a burden on limited resources. But here's the thing, capitalism thrives on selling & growth, what if there is no-one left buying, the markets are already nervous of a Global recession?

3. Will AI evolve to become an existential threat? WarGames (MGM 1983) anyone? 

4. Is there a possibility Terminator (1985) and The Matrix  (WB 1999) are already becoming self fulfilling prophesies. Welcome to the MetaVerse...

[Nathan Hildebrandt]

Firstly I don't think at this stage I would call it AI, but rather machine learning (in our field anyway). There are a lot of cool pieces of tech out there that do small portions of work, that I find really exciting, and they are being backed by a lot of money in the VC space. There will always be a place for tools like Archicad, Revit etc. There will be automation put in place to remove the mundane tasks we don't want to complete each day or speed up the process in which we do things.

Some examples.

Testfit - automatically creates optimised layouts for buildings and car parking based on requirements set by the user before they commence. Clifton and his team just got another big cash injection in the last couple of months. https://testfit.io/

Augmenta - led by Francesco has been developing a tool to automate the modelling of building services and optimise them. They are pretty early in their journey but I do see them growing and being embraced by services engineers that are fed up with Architects moving things around, and expecting the services guys to move things quickly and easily. https://www.augmenta.ai/

Hosta.ai is a really powerful tool that is using photos to determine materials and quantities for insurance claims. It is a tool I am keen to diversify into the construction and defects inspection space. https://hosta.ai/ 

ClearEdge3D - Take point cloud captures and convert most elements, mainly structure and services into model elements. There are not many automatic modelling tools out there at the moment, a lot is still manual tracing. https://www.clearedge3d.com/ 

Above are only a few of the really powerful tools that we will need to engage with moving into the future. The challenge for software like Archicad and Revit is how they integrate with all of these tools. If there is no connection then they will indeed die off. 

I am excited to move into the future. My time is valuable after being in the game for 20 years, so if I can automate cheap tasks, through automation or machine learning I really look forward to embracing it. 

[SenecaDesignLLC]

I think regardless of anything the question becomes cost. Who will be able to afford the new tech? And even that there will be many fully against a computer generated outcome. I think this is relevant but it's going to be longer to initiate and maintain for our profession than others.

Archicad can do x if an "AI* can only do x then we will all be here on the same boat in the future lol

[Karl Ottenstein]

---

@gavinNZz wrote:

AI tech is moving faster than anything we have seen before. The gains that can be made are so large there is huge money being poured into the development of AI in all industries. Architecture is no different. The video below shows where AI is already at so the next 5 yrs is going to see a big shift in the structure of Architectural practices around the world ...

https://youtu.be/h-D5-6koRf8

---

The video is a joke, right?  Seriously, what is the mocked-up "AI" doing...with what input?

Seems like smoke and mirrors and venture capital...

[Nathan Hildebrandt]

I think there are some smoke and mirrors, to be honest. There has to be a lot of prior projects completed to inform that documentation. AIDEA does have some cool APIs that they have developed to speed up the documentation process, from layout creation, view placement, automated notation and dimensioning. 

I have a contact in the VC space at Bricks and Mortar, but I find the public research that is released from Shadow Ventures is pretty accurate and a good insight into the current direction that software and hardware are taking in the AEC space.

https://shadow.vc/ 

[ryejuan]

Or will be the human species end up like in this movie Idiocracy🤔

[DGSketcher]

I do hope the dystopian future I implied above, and the natural tenacity of the human species to survive guides us on a more palatable path. Nathan's pointers to new tools and technologies is more realistic than the prospect of AI taking over the construction industry. The key point in those links is the intent of software to assist; it is still reliant on human intervention and not autonomous, thus avoiding a kick back from the prospect of mass human redundancy.

What troubles me more at this point is the current disconnect between design software developers and what happens on an average building site. It would appear that even with class leading AC, there isn't a particularly good grasp on the day to day drafting process by the software developers, and that is further frustrated by a generation of building designers with minimal site experience, who seem to lack a basic understanding of gravity or the use of building materials. Until these issues are addressed I don't expect to see a huge shift in software development, the growth area will be limited to specialist solutions.

[Patrick M]

let's not forget that this is where "AI" is at. https://techcrunch.com/2022/09/13/loab-ai-generated-horror/
We don't understand it enough to let it take over much more or our lives than a simple thought experiment at this point. Relying on "AI" (as Nathan pointed out, it's machine based learning, not actual intelligence) for something with as much liability as a $50m project is reckless at a minimum.

I do think AI CAN be a relevant tool. I think it eventually WILL be. I don't think it spells the end of Archicad, or Revit, or any other tool. You still use a pencil/pen, right? Like every day? Even though you have a computer that has all the software to replace it?

Nah; AI will not kill design/documentation software. In fact, we will need quality software even more when we start relying on computers to do the menial tasks via AI.

[Patrick M]

Yeah, i'm not 100% convinced that video IS an AI representation, TBH. it seems more likely someone just 3x sped a video walking through their model/documents/pdf and said "look what AI did!" (I may have been missing something; but there wasn't any audio... so...)