Wasn’t Elon Musk bullied at his South African school? IIRC, I read something along those lines but might be wrong.

If anyone else wants to join the email thread for a Berlin HN meetup, ping me an email - no google nor whatsapp required ;)

Having an AI generate 50 lines of code that you assume will do the job is a little different to creating a flow that you understand and maintain.

AI is really good at making everything seem so simple and easy, until it breaks. But then again, just use another AI to fix it.

So then there is an agent orchestration layer for coordinating multiple agents with multiple personalities, role playing multiple parts in a comedy called software development.

We still don’t know what they are really doing but does that matter? Does knowledge, experience, learning and failing even matter if AI is doing it for us?

It becomes a problem of organising my thoughts so that I can find them and reconnect them and bundle them together into ideas. Which later might become understanding and wisdom.

Obviously this is not nothing new, Zettelkasten is solving the exact the same problem. And through out history many collections of notes have been found and archived. These systems will benefit much from LLMs, purely for the organizational advantageous.

I have taken another approach inspired by mind maps to create my own visual PKM[1]. It has become a giant scrapbook with notes, videos and imaginary, ideas and half-finished projects, and everything in-between. The advantage is that it's a giant visual program, it's a mind map created using a visual programming environment. I'm constantly extending it with new tooling to help me keep my thoughts in order.

In the end, it's the tweeking at my scrapbook while adding to the scrapbook that then causes more tweeking until something arises that was expected but which is very much welcome.

[1]: https://blog.openmindmap.org/blog/mind-maps-for-busy-minds

HN is a place of accountants.

But once someone starts to think of deleting data, then someone else (the enemy) might also think of deleting data (the oil of the 21st century). So being preemptive and deleting first (and asking questions later) is the only logical solution. Hence CRAZY-as-a-Service will provide an easy to install Web-App for detecting thoughts on deletion.

[1] https://en.wikipedia.org/wiki/Mutually_assured_destruction

Please define viable solution. It's like a piece of string: how long is it? It's a very subjective statement "6 weeks to find a viable solution" - could just as well be 6 months.

Sure you can deliver something that might seem viable and the customer might also a cycle of updates as the viable solution becomes the solution - but how many updates are tolerated and how bad are the problems in the seventh week?

Sure we dumped waterfall for agile and now we live in a constantly updating world of "viable solutions" ...

An AI won't have written that - that would imply that AIs would be criticising their overlords and masters.

For those old enough to remember MAD (Mutually Assured Destruction), it is time to become CRAZY (Cyber Recursive Annihilation, Zeroing Yottabtyes of data) - the digital successor.

Being able to ensure that if our data is deleted, then the data of our enemies is also deleted _without_ us doing anything.

This sounds crazy but so was MAD. I wonder whether the various secret services around the world are already working on such systems and whether the next Unicorn startup could be a AI-driven predictive preemptive protection service. The idea would be to delete the data of our enemies before they can delete our data based on predictive AI.

That is, to delete our enemies data just before they think of deleting our data. And that's were the (holistic) predictive AI comes in: to predict when our enemies _thinks_ of deleting our data.

Instructive for such an endeavor would be Dr. Strangelove.

Comments URL: https://news.ycombinator.com/item?id=49096535

Points: 3

# Comments: 4

I think at the core of our arguments lies this. Turing representation, i.e., the Turing Machine is the fundamental computing engine that can represent any problem and provide a solution to that problem, in whatever n-space dimension the problem finds itself in. It is the binary representation of any problem space.

Yet no one uses it to do the real world problem solving. Why? Because its symbolism is to verbose for even the simplest solution.

Herein lies what I have come to realise: its verbosity that a 2D representation can reduce at the expensive of flexibility. As the textual representation is a trade-off, so is the visual one.

Text happens to be extremely flexible for problem representation and solution but it is less verbose than a Turing Machine and based on the self principles, so we can be assured of its universality. Yet textual representation can still become very verbose and hard to understand (hence version control and diff'ing tools).

A 2D solution can be even less verbose at the price of flexibility - a 2D representation is more bound to the problem space being addressed.

I think the win comes when a recognition arises that points to a visual representation as being less verbose and therefore more understandable because the problem space is confined and well understood that there is no longer a need for texts flexibility. This being akin to a DSL definition for the problem space.

Exactly what I am implying: using the computer to provide a better abstraction is what is needed.

I agree that text has a certain universality however as you also say, a lot folks don't work with symbols, they work with images (for example) or colours (synesthesia) or shapes perhaps.

Besides move up in dimensions is a continuation of a natural progression in computing: cogs, switchs, punch cards, keyboards, mouse, touch .... VR, AR etc. Currently we're (coders) are somewhere between keyboard and mouse while others (most folks that use smart phones) are at touch. Smart phones users are also programmers who code their device using largely visual concepts (at least I've not seen anyone set their alarm using a terminal on a smart phone!).

Now "Oh but it's not possible to represent those concepts that we work with visually, only text will do" argument doesn't really hold water since it just demonstrates a lack of imagination to find a way to use other representations for those concepts that are essential to programming. Strangely that was possible when moving from switches to keyboards, why shouldn't it be possible even further.

Recently I had the thought that in fact, textual programming (as a paradigm) is now the sole domain of AIs. I've heard so many coders say they have not coded a line of code in months/years, AI is doing all the coding for them. Hence, I think, we humans (those of us here who are humans) should step up to the next domain for coding: 2D and then 3D and explore improved abstractions for the concepts we try to represent in a textual manner (even these concepts are abstractions of the real world).

    \   |   /
      .-'-.
   -- (   ) --
      `-.-'
    /   |   \

Much like HN is full of textual description and links to images - this entire forum demonstrates the limitations of a textual visual representation.

Yes it works but it's far more verbose than an image would be. Hence an image is worth a 1000 words. And hence HN is more verbose than it needs to be. So is coding more verbose than it needs to be. That's why DSLs (Domain Specific Language) get invented: to reduce the verbosity.

And so it is with a 2D representation: it reduces the verbosity required to transport ideas and concepts. UML is another invention to reduce the verbosity and increase the understanding of the representation of complex systems. Yes UML can be represented as text (see Mermaid[1]) but the UML image is probably a lot more understandable than the Mermaid representation.

[1]: https://github.com/mermaid-js/mermaid

As a 2D representation, I've been using flow based programming (FBP)[1] and it's incarnation in Node-RED[2]. The trick here is to understand how to use FBP to create more generalistic programs, i.e., to prove it's Turing Complete.

I think it would be a mistake to just start out with a 2D approach without also taking all the learnings we have made in 1D, i.e. TDD, OOP, KISS, DRY, Functional, Message Passing, Processes etc, and porting those to a 2D representation.

[1]: https://jpaulm.github.io/fbp/

[2]: https://nodered.org/

It's surprising that there is any art at all if 1D textual representation is the ultimate way of representing concepts of the world around us. Incredible really that folks waste their time define colours for such thing as the sunshine.

Oh sorry, of course, you were talking of code - meaning that algorithm concepts are best represented in text? Again, I doubt that very much. Laziness of humans is probably the bigger reason why we're stuck with an inefficient textual representation of higher dimensional concepts.

- Birds combined with concrete block builds

- dead old people (and grumpy looking in the case of beethoven) combined with concrete blocks

- busy and abstract designs that end up being meaningless

Bland, boring & repetitive designs - what else can one expect from a technocracy.

Where is inspiration and future when europe continues to look back to the glories past?

Professionalism and being a professional means accepting the situation and getting on with the job.

After all, that's why folks get paid: to forget the pain of being stuck in a shit job, eight hours a day, fives days a week (if not more), until they're 65 - 9 to 5, to 65 - 925265, I use that as my passcode just to remind myself.

You won't be able to sell your craft, correct. That means keep your day job. If your day job requires vibe coding, then that becomes your day job. If you don't like it, quit. Live on the street and by your moral beliefs. Or else, sell your soul.

That's the stark choice that plenty of folks have faced and will face. There is nothing special about being a coder anymore - welcome to the factory floor.