It is very impressive indeed, but impressiveness is not the same as usefulness.
If important further features can’t get implemented anymore
The usefulness is pretty limited.
And usefulness further needs to be weighed against cost.
Oh I doubt it, unless you have that person with vision to interpret the results of the usability testing and turn them into a single cohesive design.
Good UX comes from someone that has deeply internalized the problems a piece of software is solving for users and the constraints on those users. Most startups do this without usability testing by doing things like sales or customer support. Anyway, IME usability testing is not the bottleneck to good UI.
I don't disagree with you that you need to have a singe cohesive design vision based on solving for users. But I think that certainly usability testing can lead to even better results and is mostly constrained by cost.
I don't think that the user you are responding to is anti-innovation, but rather points out that the usefulness of AI is oversold.
I'm using Copilot for Visual Studio at work. It is useful for me to speed some typing up using the auto-complete. On the other hand in agentic mode it fails to follow simple basic orders, and needs hand-holding to run. This might not be the most bleeding-edge setup, but the discrepancy between how it's sold and how much it actually helps for me is very real.
I think copilot is widely considered to be fairly rubbish, your description of agentic coding was also my experience prior to ~Q3 2025, but things have shifted meaningfully since then
Copilot has access to the latest models like Opus 4.6 in agentic mode as well. It's got certain quirks and I prefer a TUI myself but it isn't radically different.
It can also be a measure of other market factors. In the case of Siemens Energy the high demand for electricity generation capacity from AI certainly plays a role.
> A 2014 systematic review concluded that vitamin D supplementation does not reduce depressive symptoms overall but may have a moderate benefit for patients with clinically significant depression, though more high-quality studies were determined to be needed.
The meta-analysis cited in the article is from 2024 and specifically mentions the Shaffer et al. 2014 review cited by Wikipedia as being low quality:
> Some of the available reviews, owing to the limited number of trials and methodological biases, were of low quality (Anglin et al., 2013; Cheng et al., 2020; Li et al., 2014; Shaffer et al., 2014).
Therefore what to do? I have seen these Hacker News vitamin D ads appear every few months for the past 15 years, or so. I always seem to have a vitamin D deficiency, so it reminds me to take supplements. I take them for a few months, hoping to see a change, but I don't feel any benefit. Then, I forget to take the supplements until the next time I see an ad. How to know if they're actually doing something useful?
Remember when HN had an entire year of articles about how we should all be on microdoses of various hallucinogenic and psychoactive substances for "peak performance"?
Good times.
HN is as bad at medicine as it is at everything else.
No, I was here just for the year in which they ended COVID with vitamin D. With the weekly (n=5) articles garnered with blog posts from randos swearing by it...
If you're not in the respective fields it can be pretty difficult to distinguish good from bad research. I am not able to do so.
If you (or your close ones) don't suffer from depression, then I guess it's best to ignore it until scientific consensus has formed. That will for sure show up on wikipedia. As far as I can see as a layperson there is a lot of correlation with Vitamin D that breaks down in interventions and Vitamin D is recommended mostly for babies and elderly people. On the other hand I see Vitamin D pushed as a miracle drug not unlike Vitamin C used to some decades ago and regular reports of overdosing of supplements leading to organ failure.
If you're suffering from depression, you should talk to your doctor. They will be able to help you to weigh potential benefits with risks
Building nuclear power stations includes a lot of labor-intensive hard to automate tasks like construction. Baumol's cost disease means it's getting even more expensive: rising general productivity leads to higher wages and higher costs in fields that cannot increase productivity as much as the general economic growth. That's why it's also still cheaper in countries with access to low-cost labor.
SMRs are a try to get out of it by building more but smaller reactors. The reality is however that nuclear has an issue with scaling down. Output goes down way faster than costs and most SMR designs have outputs far greater than what initially counted as an SMR.
Investment in renewable energy already greatly outpaces investment in fossil energy. The economic decision to keep using a fossil system is a different one than having to choose a new one. There's still problems that have no economically competitive renewable solution yet, but a lot of what you are seeing is inertia.
Base load electricity is simply an economic optimisation: demand is not flat, but the cheapest electricity source might only be able to create a relatively flat output. You'll need more flexible plants to cover everything above the base load. If you have cheap gas, base load does not make any sense economically.
Even if it will make software engineering drastically more productive, it’s questionable that this will lead to unemployment. Efficiency gains translate to lower prices. Sometimes this leads to very few additional demand, as can be seen with masses of typesetters that lost their jobs. Sometimes this leads to a dramatically higher demand like you can see in the classic Jevons paradox examples of coal and light bulbs. I highly suspect software falls in the latter category
Software demand is philosophically limited by the question of "What can your computer do for you?"
You can describe that somewhat formally as:
{What your computer can do} intersect {What you want done (consciously or otherwise)}
Well a computer can technically calculate any computuable task that fits in bounded memory, that is an enormous set so its real limitations are its interfaces. In which case it can send packets, make noises, and display images.
How many human desires are things that can be solved with making noises, displaying images, and sending packets? Turns out quite a few but its not everything.
Basically I'm saying we should hope more sorts of physical interfaces come around (like VR and Robotics) so we cover more human desires. Robotics is a really general physical interface (like how ip packets are an extremely general interface) so its pretty promising if it pans out.
Personally, I find it very hard to even articulate what desires I have. I have this feeling that I might be substantially happier if I was just sitting around a campfire eating food and chatting with people instead of enjoying whatever infinite stuff a super intelligent computer and robots could do for me. At least some of the time.
The effects might be drastically different from what you would expect though. We’ve seen this with machine learning/AI again and again that what looks probable to work doesn’t work out and unexpected things work.
reply