I am interested to have end products with really good UX driven by peak performance and optimizations which blows my mind because it’s so well perfected as humans weren’t part of its process.
As of now, I don’t see any of that being the goals even let alone the result. Instead what I see is that everyone is in a race to have as much product out as possible by the same standards that applied to products made by humans, but even lesser quality.
I the last two decades we have gotten so used to average or less than average product quality that we are willing to pay 2000$ for a device running an operating system like macOS or windows that ‘works most of the time’.
I hope we have products thanks to AI that compete on reliability of software and hardware to the point that the big tech will have to reinvent their processes just to be able to justify prices.
> "I am interested to have end products with really good UX driven by peak performance and optimizations [..] Instead what I see is that everyone is in a race to have as much product out as possible by the same standards that applied to products made by humans"
AI is not equipped for this. Humans are part of the experience process, and AI doesn't "understand" individual humans better than other humans do.
UX relies very much on human psychology. That's why, despite all the efforts of scoring it, UX is not an objective metric:
1. AI doesn't know what "good UX" is, as this is judged by the sum of humans using it. --> Structured data of actual usage would be needed for AI to ingest, in a way that it can derive intent from every action and identify friction.
2. A "good UX" is not a transferrable metric. The best UX for a ticket vending machine is not applicable to video-editing software. --> AI would need to understand what to apply when, and at which point it needs to evolve its expectation to the user
3. A "good UX" is not an absolute metric. Photoshop/Premiere/... has a UX which was extended and refined over decades, it is considered good by experienced users but still overwhelmingly confusing for beginners. --> Someone would need to scope the AI for the correct target-group
All this is done today by doing extensive user-testing with actual humans. There are people experienced in this who can create better UX-designs from scratch than others can, and still we keep learning something when we put a group of humans in front of it and watch them use it.
"Intelligence" alone doesn't solve this.
I'm reminded of the story of Stephen Hawking's sister, telling how he invented a "better board game" as a child and always wanted to play it with her, which was apparently the most frustrating and boring game she ever played...
> As of now, I don’t see any of that being the goals even let alone the result. Instead what I see is that everyone is in a race to have as much product out as possible by the same standards that applied to products made by humans, but even lesser quality.
That is, roughly, how all automation works: You build the machine that does the humans job, barely, badly. Then you improve the machine. Producing shitty output on that journey is better than producing no output, if the domain allows for it. You learn more.
Judging by the recent global outage of Outlook, which, probably, in some way, affected just about all US commerce, I would say the future for software reliability is still looking grim. The good news is replacing your vendors bespoke solutions is easier than ever. Where I work, we’ve still not internalized that software does not have to be a top-down solution where everyone does the same thing. In substantial management and user education, that goes into getting everyone to do exactly the same thing. The idea that you could just ask Claude to solve your very particular problem, it solves it, and then you move on, maybe not even preserving the code, is not in that mental model
I think we are starting to see this in some areas: For example, for years desktop apps were built on top of slow and heavy cross-platform frameworks like Electron, because writing them natively was too burdensome. Now we're seeing rewrites in rust using native APIs and toolkits, resulting in lighter and faster apps.
Software engineering as a discipline is also shifting, and will continue shifting, to a process of continuous automated improvement. In the past, test suites were mostly a way to catch regressions. In future, they will be essential to provide a verified baseline of correct behaviour, and then you will have models looped in the background to continually merge performance and reliability improvements. Similarly you will have error and performance logs fed directly into agents to resolve incidents as soon as they arise.
We are in the baby steps of these changes at the moment, but I think in at most ten years, all non-hobbyist software will be built this way.
I think you have some serious rose tinted glasses about the reliability of pre 2005 software. Windows almost never randomly crashes anymore but it used to be a common occurrence in the 98 days.
> As of now, I don’t see any of that being the goals even let alone the result. Instead what I see is that everyone is in a race to have as much product out as possible
If you frequent sites like Hacker News or other spaces where projects and announcements are shared, it’s going to look like this.
My friends who don’t follow these spaces have no idea, because those high volume, average quality projects aren’t newsworthy.
The highest quality projects aren’t out here advertising that they were built with AI. Smart teams know that saying something was built with AI is not an attractive quality for consumers. It’s only the teams trying to increase their status and make more content for their resumes who are blasting out announcements about how much AI they’re using.
From here on out, you can assume that any product you use will have had AI used in the development, likely heavily. You’re going to start running into products that blow your mind with their quality, UX, and polish where teams will have quietly employed AI to a high degree. Human input still remains a significant a valuable part of the process, but high performing and attentive teams are actually the ones most able to extract the best results from AI. They’re also the least likely to advertise that fact.
All of the “made entirely by AI” projects you see are going to continue being fun experiments to see what AI can do or little exercises in people trying to advertise their AI skills so employers will hire them or give them raises. When the really great products come along, they’re not going to come with “made by AI” advertisements at all.
Why would you expect to have any other kind of quality than exactly the level people are willing to pay for?
> I hope we have products thanks to AI that compete on reliability of software and hardware
That will only work if most people select for quality and it's very well documented that people don't
Superlogical's Rex is a wonderful (if wonderfully nerdy) version of this. GREAT product. GREAT UX. Heavy use of AI to develop.
> as humans weren’t part of its process
Not even close IMO, a massive number of humans are part of this process... their collective input is what the model is!
> I the last two decades we have gotten so used to average or less than average product quality that we are willing to pay 2000$ for a device running an operating system like macOS or windows that ‘works most of the time’.
Sorry, but the complexity of these devices is much higher than anything that came before it.
What are you comparing this against, and why do you think it would be reasonable to expect higher quality at a similar cost?
The MVPR – Minimum Viable Press Release.
With "AI" coding tools the relevant metrics are (again) LOC and token burn. But in the real life the fundamental truth is still the same, code is to be minimized.
> compete on reliability of software and hardware
Something tells me that a generation of TikTok influencers aren’t going to turn around and demand quality from their computing devices.
The best you can hope for is that the Swiss cheese of OS security is patched up a little more. At least those problems make headlines.
AI isn't required. UX doesn't increase bottom line, so it's not targeted. Hiring a UX expert is a cost center, not just for their salary but the cost of changing things to appease them. Combine that with very little competition (to switch between Linux/windows/macos is not like switching from Ford to Toyota), and you have what we have today. People are prisoners and the UX is the food delivered by outsourced, lowest-bid third parties.
My solution is to have an independent ratings agency that gives OS' and individual apps a score. Then software makers could compete by advertising their scores.
My trust in these announcements: the goal seems to be to announce “AI can do this now, btw”. I see no care about delivering something that people want. The powers that be are happy to make the same thing but with some 100% AI attestation.
No human in the product making. No human in the product use.
This is why I am not worried about job prospects in software development.
Neural networks are just another abstraction layer that allows you to produce more functionality by typing fewer characters. A very smart, complicated library, if you will.
And just like all abstractions before it, all productivity gains will be gobbled up by capitalism.
Sure, you can use this productivity boost to do more. Some will, and we will praise them for it. But in a world where primary incentive is acquiring as much centeral bank bills as possible is the main KPI, productivity will be used to deliver the same amount of work (or less!), at lower cost.
Until the world in its entirety undergoes a dramatic moral and systemic shift, I sleep well at night.
Reliability is the last consideration right now. All that matters is time to market
> everyone is in a race to have as much product out as possible
That explains so much. Everyone (or at least, everyone we hear about) is using AI for faster, which leads to more, but not better.
> applied to products made by humans, but even lesser quality.
I agree. Considering these AI companies promise us everything at perfection, it is shocking how the quality isn't anywhere near perfect.
> I hope we have products thanks to AI that compete on reliability of software and hardware to the point that the big tech will have to reinvent their processes just to be able to justify prices.
Which part of the entire history of our industry (and other industries) leads you to have a hope like this?
Bro you missed the days of config.sys and autoexec.bat if you think today’s code is mid.
> because it’s so well perfected as humans weren’t part of its process
How does that make any sense? How is really good UX, peak performance and optimization driven by the absence of humans in the process?
This is complete nonsense.
> I the last two decades we have gotten so used to average or less than average product quality that we are willing to pay 2000$ for a device running an operating system like macOS or windows that ‘works most of the time’.
I think that describes the entire lifespan of consumer computing.
I’m old enough to remember Windows 98 running on a Gateway PC. That cost $2000 (without taking inflation into account!) and worked much less of the time than today’s OSes do. Modern OSes are frustrating in various ways but I think it’s easy to take for granted how stable they are compared to the past.