flux anyone?

Oy veh. This "cloud" business is getting stormy. What's next? Collaborative brain surgery?

After 28s of their 30s elevator pitch, they deign to spout What is Flux? Flux is a browser-based end-to-end electronic design tool that breaks down barriers. *Book a Demo* Er, no. That's not how it works,

And after 30s Gabe Ochoa Squarespace "The electronics design software space sorely needs good tools with good UX that are designed for a "software first" age. This is so cool!" Gag me with a spoon (I think that's a current kewl/hip expression)

Collaborative design of a circuit board is a recipe for six or so etch iterations. One person needs to be in charge.

Sorry, that's from 1982. Moon Unit Zappa.

Sounds as though they're not even mildly activated.

Cheers

Phil Hobbs

They did raise $12M to do this.

Well, they need some razzin', then. ;)

Cheers

Phil Hobbs

I keep getting emails from the founder guru, and keep telling him that his demo schematics are obviously stupid.

1. Few people in tech actually understand electricity.
2. Most people are strongly affected by peer pressure, emotionally influenced by other opinions and group concensus.

So an electronic design collaboration with high group visibility is a terrible idea.

There might be ways to fix this.

I assume this is another "Doing Hardware is just like Doing Software" motives. And they're motivated by the appeal and relative success in collaborative digital platform development. Where, the software components are highly defined structures and standardized; the programmers ("developers") use all sorts of support tools to allow for a distributed collaborative working environment. Is it miraculous? No. It is progress, in that people can join & leave the team. So labor cost can be managed. Ah hah! It can take time to get major things done. Tasks are broken down & prioritized into 'sprints', etc.

So what do you think - Can Hardware be broken down into same ways & managed like software?

I wonder how much actual invention, creating new architectures, is used in software projects, as opposed to just grunting out a lot of code. Grunting can reasonably be parallelized.

Certainly a few software structures needed real invention: internet protocols, file systems, file formats, os kernals.

An electronic design can be broken down into parts. Someone can do the power supplies, someone the real signal electronics, and other people the FPGA and uP code and PCB layout. But I think one skilled badass should be in charge.

If labor cost is an actual constraint for a project it's probably not worth doing. At least in my youth, we'd bill out to the company at something like 5x to 10x our individual burdened run rate.

Of course then we'd get sold and said run rate would now incorporate the cost of financing the acquisition.

In one acq, the host company bought it for $6B and sold it to private equity for $125M five years later. I once calculated how long it would take to burn tat many $20 bills one at a time in a burn barrel. It was longer than five years.

Meanwhile, the "standard" components help turn schedules from days to weeks, and nobody really knows if anything works until it's deployed. They try, with CI flows but somebody has to maintain that. And the developers spend about 80% of their time on "pilpul" level fine grained aesthetic choices with little emphasis on actual correctness.

So there are CVEs. And people exploiting the exploits, and people doing security theater about the exploits.

Have a good look at the bleeding edge ML offerings - they're completely dependent on being emergent phenomena. You can construct a story - derived from white papers - about how they work but you can never actually know.

Why would it be? Hardware development is largely[1] a solved problem. Software's completely corrupted by "social" this and "social" that. Get on the Reddit tech fora - it's amazing what people get paid to do these days.

Oh, dog yes.

"7 Revealing Ways AIs Fail. Neural networks can be disastrously brittle, forgetful, and surprisingly bad at math "

There's a limit on how much you can mark up the price when selling engineering. There is almost no limit on how many times you can manufacture and sell a design.

1. Invent goofy idea

1. Raise M from silly people who have too much money

2. Enjoy life until the money runs out

1. Goto 1

Who decides what psu voltage to use? 'No, I refuse to run the valves at 5v B+.' 'But this TRF reflex neutralised project shows it can be done. It'll save us another supply. And look, it even boasts near zero sensitivity!'

I did suggest that the person in charge should be skilled.

Of course. There are still plenty of opinion differences.

"TRF reflex neutralised project" is old-timeish radio, which was interesting. What's different now is that voltage gain used to be hard to come by, and now it's basically free.