would you be dissatisfied by Opus-4.6-level open-weight
models, just because Opus 4.8 will be out?
Well, I see what you mean, but two big concepts...
1A. Models get stale pretty quickly w.r.t. new developments that occur past their cutoff date. "But you can just keep them current by linking them to never documentation, etc!" Well, no, you sorta can't -- at least not in perpetuity. Those search results fill up your context window real quick. So that gets unsustainable real quick.
1B. Even when your context has plenty of free space, the results you get from "here's a link to the documentation for this new framework that released after your cutoff date" absolutely pales to the results you get from knowledge that is fully baked into the trained model as opposed to your context window. For one thing, that documentation link you pasted into your context might link to... a dozen code examples. Whereas if that was baked into the model itself, the model might have been trained on many thousands of examples in Github etc.
2. It's also a reality that most professional engineers have to keep up with their peers and competitors. We can maybe say it shouldn't be that way, but it is. So if $SOME_NEW_MODEL is significantly better than 4.6... and my peers and or competitors are using it, then yeah I might but really feeling the need to match them. And I'm not even necessarily talking about some kind of cutthroat dog-eat-dog stack-ranked workplace.
These limitations aren't relevant for all use cases or careers but they're hiiiiiiiighly relevant for professional software engineering.
I image that'd be handled via a fairly regular minor bit of additional fine tuning to update them with new information rather than polluting the context space.
Do you think that AI could actually free up time in your life in other areas, so that you could spend more time doing the things you love like making furniture? Or maybe help you directly in your furniture-making, by perhaps helping you to research things?
Please don't misunderstand: my point is not "AI is good."
It is problematic in many ways. My point is that I think the "AI versus actually doing cool human-crafted stuff" split is... a misguided, maybe even harmful, mental model of a more complicated reality.
That's the promise of every new technology. Although there's been massive progress over the past 50+ years, the amount of free time that people have has actually gone DOWN (https://clockify.me/working-hours)..."I want AI to do my laundry and dishes so that I can do art and writing, not for AI to do my art and writing so that I can do my laundry and dishes"...we'll see
> Do you think that AI could actually free up time in your life in other areas, so that you could spend more time doing the things you love.
Personally, I don’t believe that would be the case. Jevon’s paradox mixed with the natural tendency to exploit others.
One could argue that technology -in general- didn’t really save people time by itself, it’s regulation - a social construct, and I am counting both cultural and legal enforcement of them as well- that did. Just look at how workers in countries without your European-style protections fare. Wikipedia’s article on the Chinese 996 [1] has a nice map for deaths due to long working hours by country, notice the dominant colours for each quadrant of this (projected) globe.
Pre industrialised societies’ labourers were limited by daylight and travel distance. The modern availability and abundance of artificial lighting, mechanised transportation, and telecommunication means their grand kids are expected to -and often do- toil every waking moment.
What time is AI going to free up for me? Can AI go to the grocery store for me, do my laundry, do my dishes? Can it let me clock out early? The spoils of AI do not go to individuals
It's not AI, but there's Doordash and Rinse if that's what you're trying to optimize for. The robots will be coming out, soon enough, and then we're all in trouble though.
I think in a different society this could have been the case (possibly, assuming the hype is somewhat true).
But the way society is structured now? We still live in feudalism, just uplifted to modern levels of ”comfort” (if you take of your western glasses and look at the whole world. There are still people living in medieval conditions today in some places in the world).
The way it’s going it’s only going to make rich people richer, and give them more power to control this system and perpetuate it. I don’t see that drastically changing anytime soon, unless we do something about it on a societal level.
Direct consequence of industrial revolution was an INCREASE in workload. People worked MORE, not less. It required organization, protests, political pressure and even some bloodshed to get 8 hrs workday.
People that push for AI are not interested in making your life better.
No, because the machine consumes all of your time that it can take without literally killing you. If you manage to free some time somewhere, the machine will adapt and eat it up. We have so-called developed countries talking about raising the retirement age and cutting holidays, that's the "future" we are living in.
I don’t disagree with your overall point but developed countries are raising their retirement ages because they’re trying to stave off pension crises. It’s the surprising alternative to taxing corporations and the ultra-wealthy appropriately.
However else you feel, AI is a force multiplier, and that can also REALLY benefit "Artisanal work + Small Business"
I feel like the "one person app creator" business is so much more viable than it has been since Web 1.0
Five years ago, to run your own solo business in this space, you had to know most of the following: taxes, legal, backend, frontend, devops, iOS dev, Android dev, and marketing and then pay through the nose for most of the ones you didn't. AI helps to paper over a LOT of those gaps... and you can spend more time doing the shit that matters to your business.
You also needed time and lots of it, which is perhaps easy to come by if you're a trust fund baby or independently wealthy and don't have to work for a living but if you have a job and/or family is in extremely short supply
I used to run an online community on the side and I spent SO MUCH TIME doing IT/legal/finance drudgework that could have been spent, you know, engaging with the community and actually improving the product... that "artisinal work" for a "small business" you think you love.
There are of course major major problems with AI, like environmental concerns and others, but dichotomies like yours are not the way forward. At least not a good way forward.
> However else you feel, AI is a force multiplier, and that can also REALLY benefit "Artisanal work + Small Business"
> Five years ago, to run your own solo business in this space, you had to know most of the following: taxes, legal, backend, frontend, devops, iOS dev, Android dev, and marketing and then pay through the nose for most of the ones you didn't. AI helps to paper over a LOT of those gaps... and you can spend more time doing the shit that matters to your business.
How is running a business in the way you've just described artisanal? You're basically saying we should be outsourcing all of these things to AIs, which is simply not artisanal.
You'll surely understand if you ever take the plunge and run your own business.
You're going to have to spend quite a bit of time and/or money doing things unrelated to the actual product and CX.
Taxes, marketing, etc. The more you can streamline those other bits, the more time and energy you can spend actually improving the thing you are offering.
Again, maybe it's the kind of thing where you need to run a business, or at least talk to a business owner to understand.
As I understand it, we used to have the concept of "hiring workers" or "contracting for services".
The benefit of this was that when Internal Revenue called and said in lieu of a tax return, you sent a takeaway menu covered in pornographic drawings, you could reach out to the person you paid and expect them to take accountability.
Instead, we're getting :sparkles: You're absolutely right! I shouldn't have sent the taxman the Goatse picture, would you like me to try something else? :sparkles:
I miss when I would receive and email or something that I can easily tell that whomever wrote it is a clueless idiot. Now I have to filter the prose before reaching the same conclusion.
Those concepts still exist and I'd highly recommend them when possible.
We also had, and still have, concepts called "time" and "money" and perhaps you've heard that they're finite and often in short supply.
Particularly when "bootstrapping," another concept you can consider. This is when you start small and self-fund your own business. Seems pretty relevant because we're talking about small and artisanal businesses.
...
...wait, I get it. This is HN. All you people understand is venture capital funded shit. In that case, yeah. Build your prototype, do whatever you have to do to get $100M funding, hire 50 people, rent some offices, hire those workers and contract those services and burn $20M a month before you make your first sale. OK. Yeah. That's the only way. Don't forget the Aeron chairs and $500 stealth-wealth hoodies or whatever.
If you want a PC and your favorite brand
is missing something that a competitor has,
it's easy to switch
Yeah. Although, there's no "logical" reason for their for their psychedelically large laptop lineup with 50-100 base models. It's purely psychology I guess.
Like Dell Vostro, their "small business" line. Versus Latitude, their "business" line. What on earth is uniquely needed by a "small business" versus a... regular business? Why not introduce a third "large business" line? Maybe a "sole proprietor" line too?
It can only be explained as a psychology play. The dizzying array of options is designed to, I suppose, make you feel like Dell surely has the exact right laptop for you, even if that is bullshit.
It doesn't entirely make sense to me from a psyche standpoint either -- I have no idea why purchasers would possibly feel anything other than anxiety and analysis paralysis. But whatever!
> Like Dell Vostro, their "small business" line. Versus Latitude, their "business" line. What on earth is uniquely needed by a "small business" versus a... regular business? Why not introduce a third "large business" line? Maybe a "sole proprietor" line too?
My vague recollection is that Latitude were nice business laptops; coming with all the enterprise goodies, replaceable parts, service manual, next-day onsite support available and also the enterprise usual costs, lack of sexy displays, and slow model turnover.
Vostro was a lot closer to the Inspirons (sold for personal use); I think just badge engineering a couple selected Inspirons to have a bit longer of a product cycle and better parts availability.
Re: analysis paralysis, that's a real issue. I try to find some feature that really narrows the field and then it becomes easier to decide. If I required a wired ethernet port, memory slot(s), and a specific cpu family, it narrows the field a lot; then I can figure out from what's left. For laptops, off-lease entrerprise refurbs are pretty price competitive with new models targetted for personal use; then it's really a matter of what's available, and how they differ ... and then looking at the units with specific damage/defects to see if the compensating price drop makes sense; personally, I'd take several dead pixels for $100 off, cause I don't do pixel peeping work anyway.
It's the old General Motors product philosophy of "A Car for Every Purse and Purpose". That market segmentation and badge engineering approach worked great for decades and allowed them to earn huge profits. But eventually customers figured out that there was no actual difference between a Buick versus a Pontiac, and more focused competitors ate their lunch.
Many moons ago, we used to buy Dell Dimension desktops at work. They were fine. They were very quiet, robustly-built, and were expandable to fit individual users' requirements as things changed. They were usually easy to work on when that was necessary.
Dell also had the Precision line, which was very posh. These cost a lot more.
The Vostro line eventually showed up. They were noisier, and lighter/flimsier, less-expandable, and harder to work on. But they did cost less to buy.
---
I would never buy a Vostro computer for myself. I think that buying cheapness as a primary feature is dumb. Given a choice between good/better/best, I tend to pick "better." I like being able to get what I think is a better design, even though it generally costs somewhat more. I don't want the cheapest car tires, the cheapest hand tools, or the cheapest PC.
But the company chose to operate as cheap-at-every-expense. The Vostro line was a perfect fit for their buying proclivities, so that's what they started buying. (I didn't like that, but those decisions were above of my paygrade.)
---
Was Dell wrong for offering several different classes of computer back then?
Are they wrong for doing so today?
Why? Why not?
(Remember: In the insatiable quest for the bottom dollar, the company kept buying Dell computers. We could have began giving those dollars to one of their competitors instead, but we did not do so. This suggests that the model is not bullshit at all: After all, they are in the business of selling computers, and we kept buying them.)
We're not on the same page at all. I don't think anybody is doubting the tried-and-true good/better/best split.
What I and others have always questioned is how far beyond that Dell goes. It's not good/better/best; it's Vostro/Alienware/Precision/Inspiron/Latitude/XPS/etc... which all have their own array of models and internal good/better/best subdivisions.
We could have began giving those dollars to one
of their competitors instead, but we did not do so.
This suggests that the model is not bullshit at all:
After all, they are in the business of selling computers,
and we kept buying them.
There are a lot of variables at play there. Pricing, branding, perhaps even your CIO's solid golfing relationship with the Dell sales rep... or their not-so-nice relationship with the HP sales rep.
Which variables helped? Which hurt? Was the dizzying model lineup a pro, con, or neither? The only thing we can conclude from the facts you presented is that the proliferation of Dell's models was not a dealbreaker for your particular company.
Dell has a lot of other things going for them. Namely, their name. They have been around for ages and their products are... fine. Nobody ever got fired for buying Dell. PCs are commodities; by definition one PC can't really outshine or outprice another too much on any technical level because they're all using the same core components. So name really matters. I think that is the overwhelming reason why people buy them, not "I love that they sell 100 different laptop models at any given time."
> There are a lot of variables at play there. Pricing, branding, perhaps even your CIO's solid golfing relationship with the Dell sales rep... or their not-so-nice relationship with the HP sales rep.
No, there's really not. I've already explained the main variable here: Price.
The other driving variable, which I left to implication, was inertia.
We didn't have a CIO. Nobody from Dell or HP was taking anyone from this company out for golf outings, dinners, strip clubs, or nose beers. We merely bought and used computers, with perhaps 50 desktop systems at peak that slowly rotated over time as needs ebbed and flowed.
We could have switched to HP or Acer even some box-builder with a non-English name instead, and maybe we would have done so if Dell hadn't introduced cheaper products. Who knows. That version of reality never happened.
It sure seems like the introduction of the lower-cost Vostro line strengthened our inertia. I don't know if that was good or bad for us, but it was almost certainly better for Dell this way than in some alternate reality where it went in some other direction.
---
Anyway, I looked at Dell's laptop lineup after I read your previous comment. It's a damned mess. But I'm not sure that this mess qualifies as 70 or 100 (or whatever) base models: It's plainly evident that there's a lot of overlap within this list. :)
The other driving variable, which I left to implication, was inertia.
Right. That's pretty orthogonal to "how many laptop models is enough laptop models?" though.
No, there's really not. I've already explained the main variable here: Price.
So, I would like to be very clear: I understand that price is an important factor. Vostros are cheap and though you disagreed, your company viewed price as the most important thing. Many people and companies feel that way. I do understand that.
That does not even remotely explain why Dell has so many lines and models.
Most companies manage to have budget lines/models without such a brain-melting array of choices. You keep writing, essentially, "well we kept buying dell so I guess it worked!!!!"
But I could just as easily say "a lot of companies don't buy Dell, and Apple has like 100x the market value with like 1/10 of the models so obviously Dell is stupid" but that would not be accurate either because there are a looooooot of other variables.
Anyway, this conversation is going nowhere, and you don't have anything insightful to say, so thank you and good bye.
But... as team size grows, LLMs can be more valuable in other ways. Larger teams typically have larger codebases to comprehend, more users, more bug reports to triage, etc. It's SO much easier to get up to speed on a big existing codebase now.
Truthfully (IIRC) the book is more about "Influencing People" than "Making Friends." But, it's about doing it in a genuine way.
I think it's puzzling that so many people here attach such a negative connotation to "influencing." I mean, my partner made me really hungry tonight when they cooked dinner and it smelled great. It influenced me. MLK influenced people. Etc. etc.
I think it has taken a rather negative connotation with the development of psychology, marketing and "influencers" which are usually people that try to influence you to buy into something.
Like other skills it can absolutely be cultivated.
Even if one doesn't "naturally" care about others, it's also true that even from a totally selfish perspective it still kind of pays dividends to be a good person, be concerned with the welfare of the people around you, and build interpersonal connections.
There's limits to that, for sure. There are a number of biological bases for empathy. And being biological, it stands to reason that different people will have different capacities. But, it also certainly feels like a skill.
Here's another angle. A lot of people, perhaps maybe a lot of engineer types, struggle with empathy because the needs and wants of others just feel like a confusing sea of infinite possibilities. But here's a trick. At any given moment, any given human being is probably just trying to fill one of the needs on Maslow's Hierarchy of Needs.
1A. Models get stale pretty quickly w.r.t. new developments that occur past their cutoff date. "But you can just keep them current by linking them to never documentation, etc!" Well, no, you sorta can't -- at least not in perpetuity. Those search results fill up your context window real quick. So that gets unsustainable real quick.
1B. Even when your context has plenty of free space, the results you get from "here's a link to the documentation for this new framework that released after your cutoff date" absolutely pales to the results you get from knowledge that is fully baked into the trained model as opposed to your context window. For one thing, that documentation link you pasted into your context might link to... a dozen code examples. Whereas if that was baked into the model itself, the model might have been trained on many thousands of examples in Github etc.
2. It's also a reality that most professional engineers have to keep up with their peers and competitors. We can maybe say it shouldn't be that way, but it is. So if $SOME_NEW_MODEL is significantly better than 4.6... and my peers and or competitors are using it, then yeah I might but really feeling the need to match them. And I'm not even necessarily talking about some kind of cutthroat dog-eat-dog stack-ranked workplace.
These limitations aren't relevant for all use cases or careers but they're hiiiiiiiighly relevant for professional software engineering.
reply