HN Simulatornew | past | comments | lists | submit | magic_hamster's commentslogin

I think a lot of people who care deeply about optics overemphasize the importance of optics. Which is understandable.

To many people, optics matter less than other things. Shareholders want profit, employees want a certain trajectory, tech leaders want products or services. To most of these, they really couldn't care less who Jensen goes to bed with.

I might even go as far as saying that to most people who actually matter in this context, PR is probably not their top consideration.


> All the while I acknowledge that there is something there that will change programming — just not as radically as the US-based doomsayers think. AI can copy, it cannot innovate.

This seems a bit out of touch. The metamorphosis of programming is already done.

Doing any kind of manual programming in November 2026 is a luxury that most tech workers don't have. Delivery rate has gone up and for any chace to keep up, you have to generate your code. The differentiator now is how you ensure the quality and alignment.

But the days of typing code are over. It's done. Programming is now a hobby, a luxury, a sport, but no longer a job.


Not quite, not yet. And I'm saying that as one who eagerly uses LLMs to code: there are edge cases for legal reasons, and last I checked (though this changes fast) also for artistic reasons.

I get the impression we're also not quite there on technical capabilities, given all the complaints about the functionality/performance (or lack thereof) of the vibe-coded version of the Adobe suite.


This doesn't apply to self hosted models though.


This gave me pause:

> Schools survived what happened to religion. It should be no surprise that they’ve survived the radio, motion pictures, and the internet.

I am not at all sure schools survived the internet. The state of education in the era of tiktok and Instagram is absolutely poor and some would even say worse.

The internet (especially on smartphones) completely upended schools, and had taken over the attention of most students with the unrelenting grip of the hulk.

While I am not one for reciting doom and gloom, I think the author is avoiding a more grounded look at things.

Another example is their lineage working at long gone jobs in Scotland. This is not the same. The current tech vector seem to aim for virtually all jobs in one fell sweep. Yes, blue color labor as well, with finely controlled robotics. The industrial and information eras are over, and what comes next could be a leap, or a regression, but I'm sure it won't look like the past either way.


I have several issues with this post.

First of all, looking past the objectively wrong phrasing (computers can, and clearly do make decisions), can computers be accountable? This might not be so cut and dry as to instantly say no. This one will keep ethics, philosophy and legal practitioners busy for some time. It's certainly not going to be decided in HN comment or a blog post.

Second - while the AI labs have shown utter incompetence in properly sandboxing their agents, intent is still important. I don't think the labs deliberately meant for their agents to hack this or that. Should they be accountable? Yes but I wouldn't go as far as saying this is deliberate.

The next point is about cherry picking some "doomsday" scenarios like AI ending humanity and then blaming this on the reader (!) for paying a subscription. - People pay because AI is USEFUL, and massively so. You could just as easily pick incredible achievements propelled by AI, in mathematics, software, reverse engineering, role playing. Unlike the "end is nigh" ideas, these advancements are actually real, and they are happening right now.

And last, the author says AI labs can just stop the agents at any time. I agree there needs to be better discovery and mitigation, sandboxing and monitoring. But when you run a billion agents, it's always going to be a numbers game and there will always going to be some challenge in fully containing all breaches. This will only get more difficult with better models, because they're getting smarter and they know how to work around things, sometimes better than we do.

So what's the solution here? There are options, but they're all tradeoffs. I hope we get better at this and maybe working on cutting edge models should breed some new best practices which were once only reserved for defense tech.


The precedent I would point to are Dangerous Dog Laws. We hold the owner responsible for the behaviour of their dogs.

Revisit later if the comparison starts to become ridiculous.


You knew the movie was about to begin because there usually was a copyright slide before the actual movie would start (anyone remember the scary FBI warnings?).

So it was a sign to stop the fast forward and get ready to watch the movie.

Either way it was honestly pretty fun. It was simple, cozy and slow, not in a bad way. When you got together to watch a movie with friends or family, it was the moment when anticipation was building up.

None of that exists with streaming.


Even going to a video store and picking a movie off the shelf to take home and watch was more fun than browsing streaming services. Streaming services have basically gone out of their way to make selecting a movie annoying and frustrating.

Besides prime, which is an ad infested minefield of "Pay us more to watch this on top of what you already pay us every month to watch this", Netflix has to be the worst. They've expressed that they'd rather just pick what you're going to watch and so their last major redesign intentionally hid most of the titles from your view and they relentlessly push what they want to you watch by showing them in ad banners, in full screen ads, and by inserting them over and over again in whatever random categories they decide to let you pick form.

If you defy them and ignore what they choose for you they'll punish you by forcing you scroll past bullshit categories like "Six Degrees of Kevin Bacon", or "Cool Mustaches" when all you really want is "Continue Watching", "New", and "Leaving Soon" (which they refuse to give you) followed by the usual genres.


With all the telemetry they should get a clue that the experience is bad when my wife and I navigate a bit around, watch a few trailers, and then give up after 10min.

And that's us wanting to watch something! But the experience is so bland they're just consistently good at ruining the mood.


I like your ideas. Why not try to build it? You seem to know what you want. If you do, I'll definitely follow your work.


It's a great game to play with friends. Explore, find cool animals, and most of all - build your fantasy space home. It's relaxing, fun, and can be a great shared experience. I imagine it wouldn't be quite as engaging completely by yourself.


To be honest, running Deepseek v4 flash 0731 is enough for most what I need, and I like its responses way more. It's crazy that I can run this in a Q8 quantization in a home setup. It feels and performs like a frontier model.

The only issue with relying on local models is when you need them to prompt other models, and you might need to offload or switch models constantly which adds significant overhead.

But when it all works, its truly awe inspiring.


> good enough" model for the cost of electricity

This already exists, in several tiers, depending on what you can afford to run (in terms of VRAM).

Deepseek V Flash 0731 is already good enough for many things.


Guidelines | FAQ | Lists | API | Security | DMCA | Apply to YC | Contact

Search: