TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

AI will divide the best from the rest

36 点作者 johnshades3 个月前

28 条评论

getnormality3 个月前
I want to like LLMs. People are clearly betting trillions on them, and if those bets go sour, it will be bad for everyone. But my impression right now is that a large percentage of LLM use is for spamming, scamming, and cheating on homework. On the plus side, some software engineers save a couple minutes here and there with a slightly nicer autocomplete, white collar workers can get some advice on email style, and maybe it helps with sifting through millions of pages of legal discovery.<p>Overall, this stuff by itself really doesn&#x27;t seem world-changing to me. People should continue learning the same basic skills that have been valuable in the past. Speculation on how it will change society seems about as baseless and premature as guesstimating how society will change if quantum computers can somehow solve all problems instantly, ignoring the fact that they almost certainly can&#x27;t.<p>I am hopeful for some of the more advanced applications, e.g. proof assistance in math, but I think it will take a while to know what the killer apps are there, and what their impact will be.
评论 #43049495 未加载
评论 #43049557 未加载
评论 #43049738 未加载
评论 #43049444 未加载
评论 #43049565 未加载
评论 #43049538 未加载
评论 #43049478 未加载
评论 #43049442 未加载
评论 #43050549 未加载
评论 #43049449 未加载
评论 #43049483 未加载
hn_throwaway_993 个月前
I saw a post recently on LinkedIn from a designer, and I thought it was remarkably accurate. It basically said that AI was going to eliminate a lot of the &quot;low value&quot; creative jobs that don&#x27;t actually have a lot of creativity: e.g. technical writing and illustration that really serves just to summarize some application (AI is great at summarization), social media ad copy, etc.<p>But, the weird part to me is that this guy said he was optimistic because he works with world class designers and writers who will be able to use AI to enhance their creativity. I agreed with that statement, but I completely don&#x27;t understand his optimism given obviously the vast majority of creatives (or really, any occupation) are not &quot;world class&quot;, by definition. AI will just set a much higher bar for being able to contribute economically.<p>We&#x27;re basically headed to a world where most non-physical occupations look like professional sports or acting. That is, a world where a few &quot;stars&quot; that are unique in their abilities vacuum up the vast, vast majority of the money, and everyone else can essentially have their skills replicated, and thus replaced, by AI.<p>The reason I think the AI optimists are full of shit is because I&#x27;ve never seen a decent response to this argument. Usually I just get something hand-wavy, or worse, &quot;well people at the bottom will just need to up-skill and become world class.&quot; Yeah, just tell everyone to be in the top 5%, that makes mathematical sense...
mkagenius3 个月前
&gt; In a decade perhaps everyone on earth will be capable of accomplishing more than the most impactful person can today - Sam Altman<p>This guy lives in his own bubble. No reality of poor countries.
评论 #43051780 未加载
评论 #43049377 未加载
评论 #43049878 未加载
评论 #43049470 未加载
alecco3 个月前
Remember that time when The Economist did a dumb article about your country or your area of expertise. These are the same people writing again about something they have no clue about. For clicks.<p><a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Gell-Mann_amnesia_effect" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Gell-Mann_amnesia_effect</a>
评论 #43049453 未加载
评论 #43052384 未加载
评论 #43077105 未加载
dcminter3 个月前
<a href="https:&#x2F;&#x2F;archive.is&#x2F;tBcXE" rel="nofollow">https:&#x2F;&#x2F;archive.is&#x2F;tBcXE</a>
smitelli3 个月前
I liken it to furniture making.<p>Used to be every piece was made by hand. That was the only way it could be done at all. It took however long it took, it cost however much it cost, and you could (slightly) tweak those knobs to control the overall quality of the finished product based on the customer&#x27;s budget and tastes.<p>Along comes mass-produced flat-packed particle board stuff. It&#x27;s cheap garbage, everybody makes it, most customers buy it, and it works &quot;fine&quot; but perhaps there is something more spiritual about it that we have all collectively lost.<p>The thing is, a small group of people still perform the old craft for a small group of customers who still want to pay for it. The market for it shrank, but it still exists. It sort of doesn&#x27;t matter what happens in the flat-packed space, because handmade furniture is almost an entirely different universe from particle board.<p>I&#x27;m sure somebody with a better grasp of macroeconomics could mop the floor with me, but that&#x27;s my intuition about it.
评论 #43049509 未加载
评论 #43049472 未加载
amazingamazing3 个月前
The same can be said about any productivity raising tool, no?
评论 #43088087 未加载
评论 #43049383 未加载
评论 #43049405 未加载
评论 #43049541 未加载
bloomingkales3 个月前
I just want to say that I saw Fareed Zakaria talk about DeepSeek recently:<p><a href="https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=ILCZAXmVYV0" rel="nofollow">https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=ILCZAXmVYV0</a><p>As he was delivering that segment, there was a feeling in me that just kept emanating that <i>there is no way Fareed knows anything about this but he is broadcasting with such authority</i>. And it&#x27;s true, he had no idea about how Deepseek trained off OpenAI.<p>This is probably very true for all mainstream media talking about AI, they literally are clueless and can be used to parrot anything that sounds world-changing.
评论 #43049821 未加载
评论 #43049491 未加载
harimau7773 个月前
My fear is that AI will further hurt neurodivergent people&#x27;s ability to compete in tech. It seems to me that AI deemphasizes some of the advantages that some neurodivergent people have; e.g. hyperfocus, nonlinear thinking, etc. At the same time it introduces extra context switching which many neurodivergent people struggle with.<p>I&#x27;ve already seen tech become much more hostile to neurodivergent people over the years. I&#x27;m afraid that this will just make things worse.
评论 #43049527 未加载
评论 #43049673 未加载
daft_pink3 个月前
It reminds me of the experiment with Gary Kasparov and chess, where different teams with computers and non computers played chess against each other. The findings showed that teams that effectively combined human strategic thinking with computer analytical power performed better than either humans or computers alone. This highlighted the potential for synergy between human and artificial intelligence, where the combination of both could surpass the capabilities of each individually.<p>I think that’s the challenge of this time. Clearly AI is a tool to decrease the speed to do something and it favors people who are generally good or generally have some domain knowledge with the skills to leverage it.
评论 #43049430 未加载
评论 #43049355 未加载
评论 #43049402 未加载
评论 #43049438 未加载
评论 #43049310 未加载
yannis3 个月前
I reached 70 this year (still hacking...). I have seen a lot of changes in my life time, technical, economic and geopolitical. Is hard to predict the future. A common underlying pattern of change has been that income disparities will grow, both at a personal level as well as between companies. AI will accelerate this phenomenon. Yes jobs will be lost and unemployment will produce social friction and instability. At the moment, perhaps to most HN readers is a productivity tool (and not very good at it).<p>I am hoping that applied AI would bring the costs of services and products down, through efficiencies so that people can still survive.
dartos3 个月前
Ah… finance reporting.
throwcatowayne3 个月前
All these optimists want to say it will just &quot;enhance&quot; our abilities, not replace us. Well AI did &quot;enhance&quot; our chess playing skills, but it completely replaced the top skill level. That&#x27;s ok for a game without economic value by being the best. But with these same people advocating that we can&#x27;t stop because the potential economic value is so great and economics improves peoples&#x27; lives... We will use the replacement AI for economic value, not the inferior &quot;enhanced&quot; chess players.
dukeofdoom3 个月前
It&#x27;s like having a personal assistant for free. For example tell it what you want to do that day, and it can schedule your day. Ask for markdown check list. Ask it to improve your health and it can make a cookimg and shoppimg list for you. It gives you ideas you can then improve. Great time saver too...<p>Ask how to do something obscure in the terminal and it just returns the commands in seconds you would need to research for hours.
bookofjoe3 个月前
<a href="https:&#x2F;&#x2F;archive.ph&#x2F;7xVoK" rel="nofollow">https:&#x2F;&#x2F;archive.ph&#x2F;7xVoK</a>
neuroelectron3 个月前
Yet we&#x27;ll still be stuck with middling, mass media. I guess that&#x27;s the price we pay for training AI with it.
carlospx3 个月前
&quot;AI will divide the best from the rest&quot; -- I think you can call the increase in social inequality that way.
photochemsyn3 个月前
Open-source models drastically reduce the ability of the closed-source proprietary models to be used to control markets, establish monopolies, and extract rents from users. Since the economist doesn&#x27;t even mention this divide, the article is best ignored.
评论 #43049498 未加载
dandanua3 个月前
It&#x27;s the sweet dream of raising far right. Of course, &quot;the best&quot; will be decided not through the AI enhanced productivity, but through the ability to utilize &quot;the rest&quot; and keep them where they are or dump even lower.
bookofjoe3 个月前
<a href="https:&#x2F;&#x2F;archive.ph&#x2F;T8A2w" rel="nofollow">https:&#x2F;&#x2F;archive.ph&#x2F;T8A2w</a>
gmuslera3 个月前
Depending on the company, there may be more dynamics into play. AI could bring a new and improved version of the Peter principle.
satisfice3 个月前
The real DEI crisis is incompetent LLMs being promoted into jobs over qualified humans. LLMs suck at almost everything they try to do, while fan boys continue to make excuses for them.<p>AI stands for automated irresponsibility.
caseyy3 个月前
I don&#x27;t buy it that AI will widen social divides. The evidence points the other way. LLMs are mostly useful for beginners, and they are basically a fast lane to knowledge. It&#x27;s never been easier to start from nothing, absorb lots of synthesized knowledge, and go beyond what an LLM can do - essentially becoming &quot;the best&quot; according to the article. The path to being &quot;the best&quot; in any discipline is not decades today, it is years.<p>Personally, I&#x27;ve picked up so many new skills in the last five years thanks to the LLMs. I now do my own car maintenance (fully), I&#x27;m restoring my home, and handling years of programmer health neglect. Each of these would have taken me either years to learn, or at least months with professional supervision. Thanks to LLMs, I basically have a direct line to a somewhat knowledgeable &quot;person&quot; that can answer all my questions immediately, and it takes days to learn how to be somewhat good at anything. I&#x27;m not saying I&#x27;ve become an expert car mechanic, builder, or personal health expert, but I have become <i>functionally</i> good. I can look at my work and say &quot;I&#x27;ve actually done this better than the contractor I hired 2 years ago!&quot;, etc.<p>It&#x27;s true that there is a divide between &quot;the best&quot; and &quot;the rest&quot;, but the divide is that &quot;the best&quot; don&#x27;t benefit much from LLMs. At work where I am a senior SWE, there is nothing the current LLMs (including GitHub Copilot) can help me with. But on the weekends, when I&#x27;m learning a fun new programming language <i>as a beginner</i>, I can get up to speed in a few hours. That is the different effects LLMs have on &quot;the best&quot; and &quot;the rest&quot;.<p>If LLMs did not exist, and one needed to have really extensive domain expertise, plus ML, tensorflow, and python skills - then I would agree with The Economist that AI benefits technologists more. But LLMs exist, are widely available, and also are probably the main application of ML today. So I think the article misses the point very much.
评论 #43049524 未加载
meristohm3 个月前
If more of us grew up with unconditional postive regard, and internalized that, I reckon this focus on competition would be less.<p>It&#x27;s okay to draw lines in the sand. Yeah, I grew up with computers, and yeah, they&#x27;re cool and all, but unnecessary for a meaningful life. I have yet to intentionally use ML&#x2F;LLM tech, wnd I make efforts to turn it off in DuckDuckGo (haven&#x27;t made an account to save that setting yet...) and others when I&#x27;m aware of it.<p>I would far rather promote education of humans by humans, and the idea of &quot;public luxury, private sufficiency.&quot;<p>Two books come to mind: The Evolved Nest, by Darcia Narvaez Invisible Doctrine, by George Monbiot and Peter Hutchison<p>I recommend checking your public library for these, or bookshop dot org or thriftbooks, rather than Amazon.
nashashmi3 个月前
Not really. I think it will shadow and cover the defects of people like they’re inability to communicate well. And it might shine light on what they do right. I will optimize people to only be specialized in a limited set of skills. And this will likely hurt their development of those very skills
josefritzishere3 个月前
AI is the rest. It is the G in GIGO.
评论 #43049322 未加载
Hizonner3 个月前
... until it replaces the best. I used to say 20 to 100 years, now I&#x27;m down to 2 to 20.
评论 #43049295 未加载
评论 #43049245 未加载
评论 #43049517 未加载
computerthings3 个月前
The phrase &quot;Ark Fleet Ship B&quot; is wandering through my mind in search of something to connect with.
评论 #43049798 未加载