falkensmaize

I suspect casual users are MUCH more likely to either cancel their account or switch providers on a whim.

There's a lot of people who think the only thing keeping us out of a serious recession or even depression is AI investment.

I too, want to hear details about what this person did that they could be replaced completely with LLMs.

Management has made it very clear that we’re still responsible for the code we push even if the llm wrote it. So there will be no blaming Claude when things fall apart.

Thinking carefully about the details of implementation MATTERS. Even with crud apps. Getting something “built” fast isn’t and should not be the only consideration.

I can go to a junkyard and assemble the parts to build a car. It may run, but for a thousand tiny reasons it will be worse than a car built by a team of designers and engineers who have thought carefully about every aspect of its construction.

Absurdly high markups? They just released a very good laptop for $599. The Galaxy S26 Ultra is $1299. The OnePlus 15 is $999. A Dell XPS 16 with 32gb ram is over $2000.

I won’t argue that they charge a premium for memory and nvme, but I have never felt like I overpaid for my MacBooks or iPhones, in part because they last so long.

I always wondered - were the huge fines applicable because they shared the files, or because they downloaded the files? The two things were always conflated and poorly understood in media reports at the time.

Seems like it would be impossible to prove substantial damages from one individual downloading an album, because you have only lost the potential single sale. No different than a kid stealing a single CD in terms of lost revenue.

Sharing the song on Kazaa or Limewire or Napster however means that they could have potentially illicitly provided the album to thousands or even millions of potential customers, more akin to stealing a truckload or even a whole store full of cds. In that case, it does seem plausible that you could prove (or at least convince a judge/jury) significant damages more in line with the exorbitant punitive sums.

Since they “caught” you by setting up fake peers that recorded your ip when sharing, I always assumed it was the latter that actually got people in trouble.

Not if you can review the code.

I don’t see definitive evidence that there is some kind of Moore’s law for model improvement though. Just because this year’s model performs better than last year’s model doesn’t mean next year’s model will be another leap. Most of the big improvements this year seem to be around tooling - I still see Opus 4.6 (which is my daily driver at work) making lots of mistakes.

So you want LLMs to write a bunch of black box code that humans won’t be able to read and reason about easily? That will definitely end well.

Clearly the target audience for this device are the 90% of users who are going to use this to watch YouTube, talk to ChatGPT and upload photos to Insta, or whatever the kids are doing these days. It’s not designed or marketed at power users, although my past decade plus experience with Macs is that they can stretch a lot further than their specs would suggest.

The cost to train and run these things is going to lead to fewer players eventually. I suspect we end up with 2 or 3 big players in 10 years.

You could say the same thing about learning to fly a plane or play the piano. I’ve got no problem with people using llms to write useful utilities and programs for themselves, but it is leading us into a world where people are creating powerful things they don’t understand. That’s going to have consequences. We are already witnessing some of them.

But the friction and learning curve is what helps you understand what you’re doing. The friction is an integral part of the process.

I’d argue it’s more like driving yourself vs passively being driven everywhere. Remember that scene at the end of E.T.? Where Elliott and his brother steal the van, but they don’t know how to get to their destination because “I don’t know streets mom always drives!”. LLMs are mom always driving - you might recognize some landmarks after a while, but you don’t know the names of the streets to get anywhere.

Miele vacuums are better performers than Dyson, and cheaper too.

It's not really ironic at all. It's entirely possible for someone to be a well-credentialed liar/conman.

I don't know if this is the future or not, but it seems to serve no real purpose other than to enrich LLM company profits. There is real value in well designed code that has been battle tested and hardened over years of bugfixes and iteration. It's reliable, it's reusable, it's efficient and it's secure. The opposite of hastily written and poorly understood vibe code that may or may not even do what you want it to do, even while you think it's doing what you want it to do.

It turns out they're really great at building toys, cosplay gear and little plastic parts for things, but in general not that useful in most people's daily lives. Kind of like Ai.

I just tried this question today with Opus 4.6, and it got the answer right - “Drive. You’ll need the car”.

I changed only the destination however and it got it wrong. I think they are manually “fixing” these things as they get publicized.

“I need to go to the mechanic shop. It is 50 meters away. Should I walk or drive?

Walk. 50 meters is basically a one-minute stroll — no reason to deal with starting the car, driving, and parking for that distance.”

The last 16 months of layoffs are almost certainly not because of LLMs. All the cheap money went away, and suddenly tech companies have to be profitable. That means a lot of them are shedding anything not nailed down to make their quarter look better.

I wouldn’t trust “Dr” Oz to advise me on which brand of cold medicine to buy, much less the health of myself or my child.

All I really want from Apple is to continue perfecting their computers, phones and tablets to be the absolute best computing devices possible. As long as they keep iteratively improving those things I don’t care if they’re thought or innovation leaders in whatever hot new thing comes along.

And the 3rd and 4th place apps are “Freecash” (some kind of get paid to take surveys app) and the Peacock streaming app. These may be the most downloaded by rank, but we have no idea what the actual numbers are, or what period of time this ranking covers, which makes it a poor metric of popularity imho.

What I never understand is that people seem to think the conception of the idea and the syntactical nitty gritty of the code are completely independent domains. When I think about “how software works” I am at some level thinking about how the code works too, not just high level architecture. So if I no longer concern myself with the code, I really lose a lot of understanding about how the software works too.

Watching this stuff all day can literally cause you to have lifelong PTSD. I want poor people to have enough money to provide for themselves, but this is exploitative - they should get paid a LOT more to do this kind of work, the same way someone who does something physically dangerous gets paid more for the risk.

Pfft. Dario has been making nonsense fear mongering that never comes true.

Just today I had Claude Opus 4.5 try to write to a fictional Mac user account on my computer during a coding session. It was pretty weird - the name was very specific and unique enough that it was clear it was likely bleed through from training data. It wasn’t like “John Smith” or something.

That’s the kind of thing that on a large scale could be catastrophic.

“anti-ai sentiment”

No that’s a straw man, sorry. Skepticism is not the same thing as irrational rejection. It means that I don’t believe you until you’ve proven with evidence that what you’re saying is true.

The efficacy and reliability of LLMs requires proof. Ai companies are pouring extraordinary, unprecedented amounts of money into promoting the idea that their products are intelligent and trustworthy. That marketing push absolutely dwarfs the skeptical voices and that’s what makes those voices more important at the moment. If the researchers named have claims made against them that aren’t true, that should be a pretty easy thing for them to refute.

“Anti-AI extremism”? Seriously?

Where does this bizarre impulse to dogmatically defend LLM output come from? I don’t understand it.

If AI is a reliable and quality tool, that will become evident without the need to defend it - it’s got billions (trillions?) of dollars backstopping it. The skeptical pushback is WAY more important right now than the optimistic embrace.