Comment by olalonde
14 hours ago
Comments like these make me feel like we're living in different worlds. I use LLMs multiple times a day and they've significantly improved my quality of life. They are also steadily becoming more useful over time (e.g. now solving math problems).
I suppose many do live in different worlds.
I haven't found anything out of LLM's that has improved my life. It was a fun little toy but could never find a use case. But clearly, your mileage varies greatly from mine. That's cool.
I just personally don't the use in more when what I think many need is less. But that comes from essentially this point of view - “Better than a thousand hollow words is one word that brings peace.” ― Buddha
I use LLMs daily, both as chat applications and "vibe coding".
I wouldn't say it "significantly improved my life" however. Everything AI has done for me right now is a "Nice to have" but it doesn't fulfill my needs.
It’s because people value different things. I could not care less if LLMs make me push code faster to prod. Couldn’t care less if they improve my emails grammar. Couldn’t care less if they crack one unsolved math problem.
You do not care about these things in general or it's the idea that an LLM is involved that makes you not value them?
I do too, and pay $200/month, but anthropic’s margins on that revenue are negative.
What’s the long term plan? Make it up on margin? 100% tariffs on Chinese open weight models?
I don’t plan on pulling from my 401k for decades, so the long term plan is the part I care about.
Enterprises are paying API prices, which are ~9x the price of the plan for the same usage. A lot of people on the plans are not maxing them out either.
Why would Anthropic margin's be negative? Inference is practically free for them, so what costs do they have for an additional subscription?
amortized training costs
I’d love it if for once someone on here saying LLMs are some life changing apparatus would give a single example.
I can give some recent examples.
- Significantly increased my productivity as a software engineer.
- Using it daily for Chinese-English translation. Significantly better than pre-LLM translation software. Also, great at teaching grammar, nuances, etc.
- General Q&A. Like "Googling" but much faster. This is probably the most common use case for me.
> - Significantly increased my productivity as a software engineer.
This is exactly the point that keeps coming up that folks are struggling to grasp, myself included. How are you measuring this? It certainly makes me feel productive, but I'm not sure I can confidently say it has actually made me more productive. It's made the easy stuff a no-brainer (e.g. boilerplate, simple logic) and the moderate stuff really hard. Never mind the hard stuff. Vetting the code has become a whole other job on its own. The only folks I've found who confidently claim it increases productivity appear to be online (and without evidence), because no one in person is willing to claim that and show it.
8 replies →
> Significantly increased my productivity as a software engineer.
You’re going to have to define productivity as it applies to software engineering. With LLMs we’ve primarily seen the number of PRs over time being discussed as a proxy for LoC, as well as the speed of bootstrapping a small project. None of these have a known correlation with economic output. They just feel good, to the programmer, their manager, or both.
> Using it daily for Chinese-English translation. Significantly better than pre-LLM translation software. Also, great at teaching grammar, nuances, etc.
Yes dealing with language is the one area LLMs are actually designed for. But what’s the TAM for machine translation?
> General Q&A. Like "Googling" but much faster. This is probably the most common use case for me.
And now you’re missing any kind of traceability for the information that you “learn,” since it all gets spaghettified and then recombined into a pile of plausible slop with no attribution. Where before you had to do slightly more work to find the information you needed, now it’s available faster but you’re at complete mercy of literally 3 American companies plus the CCP for the accuracy of that information. Most people somehow seem happy with this arrangement.
1 reply →
My wife was diagnosed with several chronic conditions in the last year. AI tools both diagnosed her before a doctor did (which helped us find the right docs to care for her by figuring out what to look for). One of her conditions (Mast Cell Activation Syndrome) comes with a ton of dietary restrictions. Its helped us immensely in planning meals and identifing food triggers. All of this would have possible with out AI as a tool but would have led to much more pain and suffering and likely taken much longer to figure it out. It's easy to dismiss (especially given the hallucinations) but it's been legitimatly life changing over the last year
We have some exotic chicks the kids picked out, and 4 were going to die of brooder pneumonia.
An LLM correctly diagnosed it, and figure out that we could treat them with Nutri-drench Sheep Supplement, since Tractor Supply was sold out of the chicken version, and they are very similar.
Of course it then immediately recommended we use hemp bedding that would kill them a different way, but the saleswoman sanity checked all of the above,
100% survival rate.
Everyone’s thriving. Chickens would follow the medical advice again, I guess.
Sounds like you have only the saleswoman to thank, not the bot.
I used Gemini to fix my new boiler by taking various photos and asking it for help. It saved me a plumber call-out (this was a user error issue, nothing safety critical).
Gemini also told me about some obscure procedures to fix my wedding paperwork after it’d been submitted with typos.
> - Significantly increased my productivity as a software engineer
I don't understand this. It increased productivity of every developer in the western world, so it didn't really give you an advantage. Your output is more valuable, but your colleagues' output is more valuable too, and your competitors' output too, and so on. So you're doing more things at the same salary and it's not like your company or your employer is making more money than usual or awarding you more eoy bonus. If your "life-change" is "I'm writing more code" without any other advantage (and with the possible disadvantage of your role changing, or being at risk), why is it desirable?
Intrinsic motivation. I find developing software inherently enjoyable, and being able to build things faster with less friction simply makes me happy.
From experience, whenever someone asks in that particular tone and is actually provided with examples, they proceed to bend over backwards to "prove" that it's secretly not much of an improvement at all/AI psychosis/a mirage/actually harmful/<insert other substitute for "I don't like it therefore you must be wrong" reason here>.
I asked - I don't have an opinion either way. I use LLMs, I just don't find them life changing, but I would never deny that in a world with infinite data something that can do a pretty good first pass at parsing and summarizing and organizing it with little effort on the creation end is a good thing.
Some guy vibe coded a tasks app client that I really like. Not life changing but I couldn't find one that suited my needs since de-iPhoning before this one.
Does it have to be an LLM? https://www.independent.co.uk/news/health/ai-skin-cancer-nhs...
It does have to be an LLM; when people talk about AI these days, they're talking about LLMs; that's the common framing of the discussion. Linking to unrelated ML to argue that AI has benefits is disingenuous.
Immediate medical and childcare advice from LLM are pretty life changing.
Interpreting reports, avoiding drug interactions, or knowing when to seek medical care. And before people object- I can literally use the same LLM my doctor does to check these things, without waiting 2 weeks for an appointment.
I helped my parents work through bacterial culture results when my dad was hospitalized with sepsis, and had them ask their doctor for specific follow up tests.
I rebuilt my gas furnace and fixed my dishwasher with AI as an assistant.
Those aren't the fun parts tho. My favorite is touring art museums ancient historical sites with an LLM guide. It can give me a short academic essay about every artist, painting, or artifact. It can pull out details quirky stories about the history that I specifically would find interesting.
I cant recommend this enough. Its like visiting with a 10 PhD docents in art history.
How do you know it is not hallucinating those quirky stories :)?
3 replies →
Try the prompt to opus 4.8:
does "das man" know they are part of the crowd?
https://www.the-scientist.com/chatgpt-and-alphafold-help-des...
Literally saved his dog's life.
I think it is a typical example of where N% (N tends to zero) of the population GREATLY benefits AI models, while the next bracket (casual users) enjoy some benefit, but the vast majority would not feel any difference if they lost the tech tomorrow.
Let me rephrase that to you: The vast, vast majority of people, even in the western world, even the white-collar part of the population, are not whales or power users of AI models.
I use ChatGPT daily. And I never spend more than $25/month. If I lost it, it would suck, but it would not affect my life significantly. I then see people spending $100 / day on Claude Code tokens, programmers in startups / tech companies rack up thousands a month in bills. These people are literally spending 100x more than me, a casual user.
Yeah, I suspect they follow some sort of whale economics - where a relatively small userbase (in the big picture) and providing them with a huge chunk of their revenue.
But still these companies are being valued as if they're some omnipresent companies which humanity simply can't live without.
I don't know if its really improved my life at all. Sure I can put together quick and dirty single-use programs faster I guess, but I feel like losing that practice has actively made me a worse developer.
Do you use Grok multiple times per day? Is Grok solving Erdos problems?
> Do you use Grok multiple times per day?
No body who has a choice is using Grok
> Is Grok solving Erdos problems?
Mēh! At a slower rate than models a fraction of the price
> No body who has a choice is using Grok
The Grok app had over 100 million downloads in 2025, over 60 million active users, and generated $350 million in revenue. That’s a lot of people being forced to use it.
2 replies →
Comments like these make me feel like AI is a computer in the hands of a monkey, and that too the computer which is unreliable.
Even if they are, it still doesn't justify the ridiculous levels of overvaluation. They are not essentials and their consumer demand is extremely elastic.
So, let's see. LLMs made my overall coding output significantly faster, even factoring in review time and tech debt. My employer should technically benefit from this, but it doesn't really, because all its competitors use the same AIs and all their engineers increased their throughput in a similar way. So I'm not sure that I, my colleagues or the whole segment I work in really benefited from AI in any measurable way.
Your clients and your competitors' clients did benefit from this overall faster coding output though.
Eventually your employer benefitted too, from more & happier paying customers.
Finally you indirectly befitted as well - through continued employment, salary and bonuses and stock (if you own any).
But that didn't happen in reality, did it?
Clients - I don't see anyone delighted that apps are better, or cheaper, or more secure. If anything, I see more enshittification, more half-baked ideas and more fear that security is worse now that we let AIs write almost all code.
Employers - They didn't really sell more or expanded their customer base. They would have, if they had the exclusive advantage, but now everyone has AI. They can cram more features in their software quicker, but so can their competitors, and AI is not magically opening any untapped market. If anything, everyone is now doing the same thing - trying to get their software on the AI train, with mediocre results so far.
You - did you benefit really? The job market is shit due to the death of ZIRP, the nature of the job itself is changing and there's a lot of uncertainty around. If anything, employees are now laid off more, not less, and salary and bonuses are not increased in any measurable way.
It looks like to me that we have to dance this particular dance because if we don't do we're left behind. That's fine, it happens every now and then. It might even be that in the future we will have tangible advantages from LLMs - better automated health care, better learning opportunities come to mind. That has to be demonstrated. But now, in year 2026, what's one advantage of AI? Having less and pricier RAM? Being able (and expected) to write more code in less time?
Another generic, useless comment which absolutely refuses to mention any particular aspect, eg how your life has been improved. It's so tiring
And when pushed all we get is another teaser of "Significantly increased my productivity"