Live data from Hacker News

What could have been

coppolaemilio.com

91–100 of 127 posts

Re: What could have been

#91

> What could have been if instead of spending so much energy and resources on developing “AI features” we focused on making our existing technology better? This is a bit like the question "what if we spent our time developing technology to help people rather than developing weapons for war?" The answer is that, the only reason you were able to get so many people working on the same thing at once, was because of the p…

> instead of building the Great Pyramids, those thousands of workers (likely slaves) could have all individually spent that time building homes for their families. But, those homes wouldn't still be around and remembered millenia later. They would have been better off. Those pyramids are epitomes of white elephants.

Consider the income from tourists coming to see the pyramids. People traveled to Giza for this purpose for millennia, too.

This was not the original intent of the construction though.

Re: What could have been

#92
Greenfield development is infinitely easier than brownfield.

It's damn hard work to dig in, uncover what's wrong, and fix something broken - especially if someone's workflow depends on the breakage.

Flashy AI features get attention and even if they piss you off, they make you believe the thing is fresh. Sorry but you're human.

Re: What could have been

#93
post #57

More than anything, I want to get back the era when the users were the customers, not the product.

Very few people are willing to pay cash for the things that they currently pay for in attention and data.

Re: What could have been

#94
post #61

>There isn’t a single day where I don’t have to deal with software that’s broken but no one cares to fix Since when does this have anything to do with AI? Commercial/enterprise software has always been this way. If it's not going to cost the company in some measurable way issues can get ignored for years. This kind of stuff was occurring before the internet exists. It boomed with the massive growth of personal comput…

Thought exercise: has any of the money Apple has spent integrating AI features produced as much customer good-will as fixing iOS text entry would? One reason for paying attention to quality is that if you don't, over time it tarnishes your brand and makes it easier for competitors to start cutting into your core business.

Re: What could have been

#95
Humans are fundamentally irrational. Not devoid of rationality, but not limited by it. Many social phenomena are downstream from that fact.

Humans have fashions. If something is considered cool, many people start doing that thing, because it automatically gives them a bit of appreciation from most other people. It is often rational to follow a fashion and reap the social benefits it brings.

People are bad at estimating probabilities. They heavily discount the future, and want everything now, hence FOMO. At the same time, they tend to believe in glowing future prospects uncritically, because it helps build social cohesion and power structures.

This is why fads periodically flush all over our industry, and our society, and the whole civilization. And again, it becomes rational to follow the trend and ride the wave. Say the magic word (OOP, XML, Agile, Social, Mobile, Cloud, SaaS, ML, more to come), and it becomes easier to get a job, press coverage, conference invites, investments.

Then the bubble deflates, the useful parts remain (often quite a bit), the fascination, hype, attention, and FOMO find a new worthy object.

So companies add "AI features" partly because it's cool (news coverage, promotions), partly because of the FOMO (uncertainty is high, but what if we'd be missing a billion-dollar opportunity?), partly because of social cohesion (following fashion is natural, being a contrarian may be respectable, but looking ignorant is unacceptable). It's not about carefully calculated material returns on a a carefully measured investment. It may look inane, but it's not always stupidity, much like sacrificing some far-future perspectives in exchange of stock growing this quarter is not about stupidity.

Re: What could have been

#96
post #6

This raises an interesting question. The amount of money that's been spent on AI related investments over the past 2-5 years really has been astonishing - like single digit percentage points of GDP astonishing. I think it's clear to at there are productivity boosts to be had from applying this technology to fields like programming. If you completely disagree with that statement I have a hunch that nothing could convi…

railways only lost investor money because everyone was investing in a national Monopoly, so the when we did get the Monopoly everyone else lost everything. sounds like a skill problem. plenty of value was created and remain in use for decades, completely different from Slop today.

Not to mention that rail only got better as more was built out. With LLMs the more you allow them to create, to scrape, and to replace deterministic platforms that can do the same thing better and faster - the further down the rabbit hole we all go.

I look around and the only people that are shilling for AI seem to be selling it. There are those that are also in a bubble and that's all they hear day in and out. We keep hearing how far the 'intelligence' of these models has come (models aren't intelligent). There are some low hanging fruit edge cases, but just again today I spent an extra hour thinking I could shortcut a PoC by having LLMs bang out the framework. I leveraged all the latest versions of Opus, Kimi, GLM and Grok. For a very specific ask (happened to be building a quick testing setup for PaddleOCR) none of them got it right. Even when asking for very specific aspects of the solution I had in mind Opus was off the rails and "optimizing" within a turn or two.

I probably ended up using about 20% of the structure it gave me - but I could have easily gone back to another project that I've done where that framework actually had more thought put into it.

I really wish the state of the art was better. I don't use LLMs for searching much as I believe it's a waste of resources. But the polarization from the spin pieces by C-levels on top of the poor performance by general models for very specific asks looks nothing like the age of rail.

Do I believe that there are good use cases for small targeted models built on rich training data? I do. But that's not the look and feel from most of what we're seeing out there today. The bulk of it is prompt engineering on top of general models. And the AI slop from the frontier players is so recognizable and overused now that I can't believe anyone still isn't looking at any of this and immediately second guessing the validity. And these are not hallucinations we're seeing because these LLMs are not intelligent. They lack cognition - they are not truly thinking or reasoning.

Again - if LLMs were capable of mass replacement of workers today OpenAI wouldn't be selling anyone a $20/month subscription, or even a $200 one. They'd be selling directly to those C-levels the utopia of white collar replacements that doesn't exist today.

Re: What could have been

#97
post #46

Earlier quoted context omitted.

[flagged]

By all means; I also work at a start-up. That doesn't mean that everyone here does, or is interested in doing so, or will have the necessary background. All I ask of you is to present information in the spirit of the XKCD 10,000: https://xkcd.com/1053/

[flagged]

Re: What could have been

#98
TFA kind of assumes that the companies involved would have improved their software in a world in which those resources weren't spent on AI. Since much software contained long-unfixed bugs well before the GenAI boom, I'm not convinced.

Re: What could have been

#99
I've been using an app recently that added a bunch of AI features, but the basic search is still slow and often doesn't work. Every time I open it, I kind of brace myself, but it still disappoints me.

It feels like more and more products are focused on looking impressive, when all I really want is for the everyday features to just work well.

Re: What could have been

#100

While I'm somewhat sympathetic to this view, there's another angle here too. The largesse of investment on a vague idea means that lots of other ideas get funding, incidentally. Every VC pitch is about some ground-breaking tech or unassailable moat that will be built around a massive SAM; in reality early traction is all about solving that annoying and stupid problem your customers hate doing but that you can do for…

It would be much better if we invested in meaningful things directly. So much time and effort is being put into making things AI shaped for investors. The elephant in the room is that capital would likely be better directed if it was less concentrated.

If a million families each has a $1,000 to invest in new business, how would you envision the money to be invested collectively? what would be the process?
Post reply on HN