I recently watched a new show called Prime Target. Watching it frustrated me as some aspects just didn't hold together, or were totally overstated. The wife would get frustrated with me in turn when I would point these out.
"How do you have a problem with this, but Marvel or Star Wars is fine". I wasn't sure how to answer that at first, but I think your comment solidifies it. I can accept ridiculous and unrealistic scenarios as long as they are fun, and held together within an equally ridiculous world.
Theres a fine line between something that is clearly a dramatisation being frustrating or just fun.
I read Wuxia (chinese cultivation novels) which can go on for over thousand chapters. And often the plot is repetitive and thin, but I like it over cheap dramatisation that can be solved if the two parties decided to talk to each other (when there's no other reason that prevent them other than not wanting to). I'd take talking and not agreeing or being powerless over not talking and creating misunderstanding every day. Especially when the plot is all about not creating the chance to talk.
> "How do you have a problem with this, but Marvel or Star Wars is fine"
Marvel and Star Wars aren't fine--the writing is execrable whenever they aren't outright plagiarizing something else.
Star Wars was schlock meant to sell toys that just happened to become huge. Marvel is pretty much just straight up garbage across the board (some characters are interesting--the stories and world though are pretty uniformly trope-ridden crap).
Modern movies also have the problem that a lot of their revenue comes from overseas--China in particular. They can't risk having writing that is either too subtle for a foreign audience or cover themes that might get them banned by the government.
Thus we get Michael Bay syndrome--spectacle after spectacle and the minimum writing necessary to connect them.
(To be honest--this is nothing new to Hollywood--Michael Bay can trace his roots the whole way back through "Towering Inferno" to "Noah's Ark", etc.)
Haha, I love this. Strong opinions strongly held. Obviously you're not wrong, and I'm not arguing against this (mostly because I agree), but I do also think _some_ of the Marvel work can be quite....inspired (not sure thats too strong a stance, hear me out!), mostly around the animated medium. The What If series has been good, and I've loved the Spiderman animations, very similar to the recent TMNT movies. Just want to add the new Transformers animated movie was very enjoyable, which I wasn't expecting. So yeah, the Michael Bay-esque style gets old quickly in real life movies, but it can work really well when animated.
I'm sure alot of this can be traced back to Manga roots.
I was thinking more along the lines of simply the Hollywood movies where the dynamics are set by the enormous market forces which basically filters against anything which isn't lowest common denominator.
When you get down to comic books and animations, the market forces aren't quite so vicious and your "10% okay vs 90% of everything is crap" doesn't get so filtered out. Consequently, you get more of the standard curve--some stuff is bad, most stuff is average, some stuff is good, and a few gems poke up every now and then.
This is terrifying. Even though they acknowledge the issues with hallucinations/errors, that is going to be completely overlooked by everyone using this, and then injecting the outputs into their own powerpoints.
Management Consulting was bad enough before the ability to mass produce these graphs and stats on a whim. At least there was some understanding behind the scenes of where the numbers came from, and sources would/could be provided.
The more powerful these tools become, the more prevelant this effect of seepage will become.
Either you care about being correct or you don't. If you don't care then it doesn't matter whether you made it up or the AI did. If you care then you'll fact check before publishing. I don't see why this changes.
When things are easy, you’re going to take the easy path even if it means quality goes down. It’s about trade offs. If you had to do it yourself, perhaps quality would have been higher because you had no other choice.
Lots of kids don’t want to do homework. That said, previously many would because there wasn’t another choice. But now they can just ask ChatGPT for the answers they’ll write that down verbatim with zero learning taking place.
Caring isn’t a binary thing or works in isolation.
Because maybe you want to, but you have a boss breathing down your neck and KPIs to meet and you haven't slept properly in days and just need a win, so you get the AI to put together some impressive looking graphs and stats that will look impressive in that client showcase thats due in a few hours.
Things aren't quite so black and white in reality.
I mean those same conditions already just lead the human to cutting corners and making stuff up themselves. You're describing the problem where bad incentives/conditions lead to sloppy work, that happens with or without AI
Catching errors/validating work is obviously a different process when they're coming from an AI vs a human, but I don't see how it's fundamentally that different here. If the outputs are heavily cited then that might go someway into being able to more easily catch and correct slip-ups
Making it easier and cheaper to cut corners and make stuff up will result in more cut corners and more made up stuff. That's not good.
Same problem I have with code models, honestly. We already have way too much boilerplate and bad code; machines to generate more boilerplate and bad code aren't going to help.
Yep, I agree with this to some extent, but I think the difference in the future is all that stress will be bypassed and people will reach for the AI from the start.
Previously there was alot of stress/pressure which might or might not have led to sloppy work (some consultants are of a high quality). With this, there will be no stress which will (always?) lead to sloppy work. Perhaps there's an argument for the high quality consultants using the tools to produce accurate and high quality work. There will obviously be a sliding scale here. Time will tell.
I'd wager the end result will be sloppy work, at scale :-)
I think a lot about how differentiating facts and quality content is like differentiating signal from noise in electronics. The signal to noise ratio on many online platforms was already quite low. Tools like this will absolutely add more noise, and arguably the nature of the tools themselves make it harder to separate the noise.
I think this is a real problem for these AI tools. If you can’t separate the signal from the noise, it doesn’t provide any real value, like an out of range FM radio station.
don't you think the problem of checking for correctness then becomes more insidious then? we now can generate hundreds of reports that look very professional on the surface. the usual things that would tip you off that this person was careless aren't there -- typos, poor sentence construction, missing references. just more noise to pick through for signal
It's possible that you care, but the person next to you doesn't, and external pressures force you to keep up with the person who's willing to shovel AI slop. Most of us don't have a complete luxury of the moral high ground at our jobs.
Maybe this would make sense if you saw the whole world as "kids" that you had to protect. As an adult who lives in an adult world, I would like adults to have access to metal tools and not just foam ones.
If 20% of people don't care about being correct, the rest of everyone can deal with that. If 80% of people don't care about being correct, the rest of us will not be able to deal with that.
Same thing as misinformation. A sufficient quantitative difference becomes a qualitative difference at some point.
> If you care then you'll fact check before publishing.
Doing a proper fact check is as much work as doing the entire research by hand, and therefore, this system is useless to anyone who cares about the result being correct.
> I don't see why this changes.
And because of the above this system should not exist.
Then the hallucinated research is published in an article which is then cited by other AI research, continuing the push the false information until it’s hard to know where the lie started.
The majority of human written consultant reports are already complete rubbish. Low accuracy, low signal-to-noise, generic platitudes in a quantity-over-quality format.
LLMs are innoculating people to this kind of low information value content.
People who produce LLM quality output, are now being accused of using LLMs, and can no longer pretend to be adding value.
The result of this is going to be higher quality expectations from consultants and a shaking out of people who produce word vommit rather than accurate, insightful, contextually relevent information.
I don't think so. Instead of SEO, I think we'll soon see 'LLMO' dominating such uses, where LLM summaries are reshaped by vendors and etailers to misrepresent facts in ways that favor them over others.
I suspect this can be done simply by poisoning a query with supplemental suggestions of sources to use in a RAG, many of which don't even have to be publicly available but are made accessible to the LLM (perhaps by submitting hidden URLs that mislead the summary along with the query).
But even after such a practice is uncovered and roundly maligned, that won't stop the infinite supply of net con men from continuing to inject their poisons into the background that drives deep research, so long as the LLM maker doesn't actively oppose this practice actively and publicly -- which none of them have been willing to do with any other LLM operational details so far.
In fact, I predict that if a LLM summary like DR's does NOT soon provide references to the sources of the facts it relies on, in no time users will disregard such summaries to be yet more uselessly unreliable pfaff from yet another net disreputable -- as we do with search engine summaries now.
let's be real for a sec, i've done consulting and have a lot of friends who still do. three times in four, your mckinsey report isn't super well-founded in reality and involves a lot of guesstimation.
Hi tmnvdb, since you seem to love these super smart LLMs I thought it would be fun to have openais o3-mini-high analyze your recent comments in contrast to the Hacker News Comment Guidelines. Here is the output it gave me, hope it helps you:
------
Hey, I've noticed a few things in your style that are both strengths and opportunities for improvement:
Strengths:
- You clearly have deep knowledge and back up your points with solid data and examples.
- Your confidence and detailed analysis make your arguments compelling.
Opportunities:
- At times, your tone can feel a bit combative, which might shut down conversation.
- Focusing on critiquing ideas rather than questioning someone's honesty can help keep the discussion constructive.
- A clearer structure in longer posts could make your points even more accessible.
Overall, your passion and expertise shine through—tweaking the tone a bit might help foster even more productive debates.
------
Just reply here if you want the full 500+ words analysis that goes into more detail.
I don't get the discussions around side project and they're ML engineers, not security experts. Why are you excusing a company for a serious security leak.
If you're releasing a major project into the wild, expect serious attention and have the money, you get third parties involved to test for these things before you launch.
Now can we get back to discussing the real conspiracy theories. This is clearly a disinformation piece by BigAI to add FUD around the Chinese challenger :-)
> I don't get the discussions around side project and they're ML engineers, not security experts. Why are you excusing a company for a serious security leak.
No one is here as far as I can tell. But if you've ever been a software engineer who is required to work with someone purely from an ML lab and/or academia, you'll quickly discover that "principled software engineering" just isn't really something they consider an important facet of software. This is partly due to culture in academia, general inexperience (in the software industry) and deeply complicated/mathematical code really only needing to be read by other researchers who already "get it", to a degree.
Not an excuse but rather an explanation for _why_ such an otherwise impressive team might make a mistake like that.
Yeah, you're right, I was conflating the excusing bit.
I haven't worked with serious ML engineers, but having worked in large webdev there's usually a team involved in these projects, including senior none devs who would ensure the correct checks and balances are in place before go live. Does this not happen in ML projects? (of course there are always exceptions and unknowns that will slip through, I don't know if that was the case here, or something else)
> Yeah, you're right, I was conflating the excusing bit.
No worries. :)
> Does this not happen in ML projects?
Consistently? No. At the level of e.g. OpenAI/Anthropic? It is mandatory. These are not just research labs, they're product (ChatGPT, Claude) companies. These American companies have done a reasonable job at hiring for all sorts of skillsets to keep things well rounded.
Perhaps DeepSeek hasn't learned this lesson yet... Or, well - it could be far more complicated than that. Speculating is only so useful with so little information.
Ha, I'm just reading this after tearing my hair out after trying to get a basic omniauth google login working in Rails. I was hitting CORS issues, for hours. I'm new to rails, but had read how productive it makes you. Well, thanks to one of their productivity features (Turbo) I was ready to throw the whole thing out.
Yes, I know, this wasn't an oauth2 issue directly. /rant :-)
Although the ability to censor is somewhat interesting and important to understand at a technical level, the amount of pearl clutching and fear mongering going around in traditional media about DeepSeek is extraordinary.
Even so called independent publications are showing extreme bias. Not once does the concept or word "hallucination" appear here, now it's "misinformation". And all these concerns about submitting personal information, while good advice, seem strangely targeted at DeepSeek, rather than any online service.
Perhaps a minor point but hallucination was never a good description for errors produced by the model - all responses, correct or incorrect, are in essence hallucinations.
As a motorcyclist, counter-steering is a very pronounced and useful feature. I've tried employing the same technique on my bicycle and it had no effect. I'd be keen to understand others experiences of countersteering on a bicycle.
Can you clarify what is it exactly that you consider counter-steering?
My understanding is that it means "briefly turning the handle-bars to point the front wheel in the opposite direction of the intended turn, causing the vehicle to start tipping over in the direction of intended turn", which is exactly how you steer both motorcycle and bicycle.
I too am a motorcyclist (and now, mostly, a cyclist) and think I may have misspoke (steering vs counter-steering).
When I learned to ride a motorcycle I was taught to push the handle bars with the hand on the side I wanted to turn (so, if trying to turn right, push with the right hand); this causes the bike to "fall" on the side of the turn, and follow the turn.
This is what I meant by "counter-steering" but 1/ it only works at relatively high speeds (above, say, 20 mph, which isn't high on a motorcycle, but pretty high on a bike) and 2/ it doesn't "prevent" the bike from falling, it makes it fall, which is what we want.
Following the same principle, staying upright on a bicycle involves steering, not counter-steering: when a bike starts falling to one side, turning the wheel to that side makes it want to fall to the other side; and if done fast enough and often enough (as all riders to), maintain the bike upright.
> This is what I meant by "counter-steering" but 1/ it only works at relatively high speeds (above, say, 20 mph, which isn't high on a motorcycle, but pretty high on a bike)
Balancing is the easy part. Progressing from stand still, to pedalling, while maintaining balance proved much harder (for my child). However, once mastered, the transition to confident rider was fast, I'm sure mostly thanks to having started on a balance bike early, and never having an interest in those scooters that every other kid seems to love (seemingly at the expense of learning to ride a proper bike)
This is great timing. I recently purchased a micro:bit for learning with my young daughter (who loves it) and found I was very quickly out of my depth with even the most rudimentary customisation for the board.
My draws have now exploded with breadboards, alligator clips, jump wires, LCDs and various other electrical components and I'm in desperate need of understanding the fundamentals of how all these things work.
There's something magical and addictive about being able to control your own hardware components from your own code though. We've had great joy from simply lighting up LEDs and programming our IR receiver.
Combinatorial can be quickly calculated primarily using factorials. 4 possible options, each where you're picking all 4 exactly once is 4!. The reasoning is pretty intuitive, when you start selecting there are four options, when you go to pick the next one there are 3 left in the pool, then 2, and finally 1. This turns into 4 * 3 * 2 * 1 = 24.
I want to not like this comment, but I think you are right! There's a reason people like to say your watch has more compute power than the computers it took to put man on the moon.
"How do you have a problem with this, but Marvel or Star Wars is fine". I wasn't sure how to answer that at first, but I think your comment solidifies it. I can accept ridiculous and unrealistic scenarios as long as they are fun, and held together within an equally ridiculous world.
Theres a fine line between something that is clearly a dramatisation being frustrating or just fun.
reply