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。

Revenge of the Junior Developer

68 点作者 ado__dev大约 2 个月前

18 条评论

disambiguation大约 2 个月前
Everyone: telling me how great AI is.<p>No one: making anything great with AI.<p>Sourcegraph: an AI company, routinely promoting their LLM-optimism blogposts to HN, perpetuating the hype cycle their business model depends on.
评论 #43448107 未加载
评论 #43447766 未加载
评论 #43451962 未加载
评论 #43447879 未加载
评论 #43468014 未加载
评论 #43447768 未加载
nickysielicki大约 2 个月前
I’m usually one of the people complaining about hype cycles, and it’s usually been correct to be pessimistic about them.<p>But in this particular case I have to think a lot of people just haven’t tried it in its best form. No, not a local model on your MacBook. No, not the web interface on the free plan. Go lay down $300 into API credits, spend a weekend (or maybe two) fully setting up aider, really give it a shot. It’s ultimately a pretty small amount of money when it comes to figuring out whether the people who are telling you there’s an existential risk to your livelihood on the horizon are onto something, don’t you think?
评论 #43447836 未加载
评论 #43447609 未加载
评论 #43448329 未加载
评论 #43447673 未加载
djha-skin大约 2 个月前
&gt; I have bad news: Code completions were very popular a year ago, a time that now feels like a distant prequel. But they are now the AI equivalent of “dead man walking.”<p>I disagree. I view this as a &quot;machine guns versus heat sinking missiles from the 70s&quot; dichotomy. Sure, using missiles is faster. However, sometimes you&#x27;re too close for missiles. Also, machine gun rounds are way cheaper than missiles. However, when they first came out, missiles were viewed as the future. For a while, fighter jets were made without machine guns, but they added them back later because they decided they needed both.<p>Sometimes I find I want to drill down and edit what Claude generated. In that case, copilot is still really nice.<p>With regard to ai assisted coding: the more you know what you&#x27;re doing, the more you know the code base, the better result you&#x27;ll get. To me it feels like a rototiller or some other power tool. It plows soil way faster than you can and is self propelled, but it isn&#x27;t self directed. Using it still requires planning and it&#x27;s expensive to run. While using the tool, you must micromanage its direction, constantly giving it haptic feedback from the hands, or it goes off course.<p>A rototiller could be compared to a hired hand plowing himself, I guess, but there&#x27;s way less micromanagement with a hired hand vs a rototiller.<p>Kind of like horses and cars. Horses can get you home if you&#x27;re drunk. Cars can&#x27;t.<p>The proper use of AI agentic tools is like operating heavy machinery. Juniors can really hurt themselves with it but seniors can do a lot of good. The analogy goes further: sometimes you need to get out of the backhoe and dig with smaller tools like jackhammers or just shovels. The jackhammer is like copilot -- a mid-grade power tool -- and Claude code is like the backhoe. Clunky, crude, but can get massive amounts done quickly, if that&#x27;s what&#x27;s needed.
评论 #43447657 未加载
tyleo大约 2 个月前
I love how new technology becomes like religion. It develops both cult followers and critics.<p>In that lens think the AI cult is more right than the Crypto cult. At least I can use it to do something tangible right now while crypto is still pretty useless after many years.<p>In some sense I think these technologies need the cults and the critics though. It’s good to have people push new things forwards even if everyone isn’t along for the ride. It’s also good to have a counter side poke holes. I think the world is better with both optimists charting new paths forwards and pessimists making sure they don’t walk right off a cliff.
评论 #43447706 未加载
评论 #43447749 未加载
monsieurbanana大约 2 个月前
Even if I would agree with everything the article says, I have no idea how the author gets to the conclusion that junior developers will prevail because they are faster at adopting LLMs.<p>Didn&#x27;t he just made a point about how fast the situation is evolving? I had some FOMO about ai last year, not anymore. I don&#x27;t care that I don&#x27;t have time to fully explore the current LLM state of the art, because in a month it will be obsolete. I&#x27;m happy waiting until it settles down.<p>And if their scenario ends up happening, and you can basically multiply a dev&#x27;s productivity by N by paying N x K dollarinos, why would you chose a junior dev? It&#x27;s cheaper, but sometimes a junior dev doesn&#x27;t take longer to arrive at a solution, it never does (same for senior devs, don&#x27;t get me wrong, but it happens less often).
评论 #43448929 未加载
davydm大约 2 个月前
the only good part was the joke about &quot;vibecoding&quot; (<i>shudder</i> what a stupid term) being like a fart and attracting flies... ok investors<p>still, this &quot;ai code tools will deprecate real programming&quot; bullshit will one day be laughed at just like how most of us laugh at shitcoin maniacs<p>it just takes a lot of people way too long to learn
评论 #43447608 未加载
评论 #43447697 未加载
6510大约 2 个月前
I had this rather comical picture where developers finally get to experience what it is like to have someone write software for you. You get sort of what you asked for but it is obviously wrong to even the most novice user. You then change the requirements and get something entirely different but equally wrong or worse... and a new invoice. hahaha It gets more funny the more I think about it.
shove大约 2 个月前
I read the whole thing, top to bottom, and by the 80% mark, I still wasn’t sure whether the piece was written in earnest or the sharpest piece of satire I’ve seen in a decade.
评论 #43461413 未加载
jsdalton大约 2 个月前
Much of this post was spot on — but the blind spots are highly problematic.<p>In this agentic AI utopia of six months from now:<p>* Why would developers — especially junior developers — be assigned oversight of the AI clusters? This sounds more like an engineering management role that’s very hands on. This makes sense because the skill set required for the desired outcomes is no longer “how do I write code that makes these computers work correcty” and rather “what’s the best solution for our customers and&#x2F;or business in this problem space.” Higher order thinking, expertise in the domain, and dare I say wisdom are more valuable than knowing the intricacies of React hooks.<p>* Economically speaking what are all these companies doing with all this code? Code is still a liability, not an asset. Mere humans writing code faster than they comprehend the problem space is already a problem and the brave new world described here makes this problem worse not better. In particular here, there’s no longer an economic “moat” to build a business off of if everything can be “solved” in a day with a swarm of AI agents.<p>* I wonder about the ongoing term scaling of these approaches. The trade off seems to be extremely fast productivity at the start which falls off a cliff as the product matures and grows. It’s like a building that can be constructed in a day up to a few floors but quickly hits an upper limit as your ability to build _on top of_ the foundational layer of poorly understood garbage.<p>* Heaven help the ops &#x2F; infrastructure folks who have to run this garbage and deal with issues at scale.<p>Btw I don’t reject everything in this post — these tools are indeed powerful and compelling and the trendlines are undeniable.
xyzzy9563大约 2 个月前
I think a lot of people are going to be surprised at how fast &quot;vibe coding&quot; with agents replaces a lot of traditional software engineering. Especially for non-critical software, but eventually where safety matters too since AI can generate tons of test cases.
评论 #43447695 未加载
评论 #43449130 未加载
mdaniel大约 2 个月前
I still have to get used to the fact that a (sourcegraph.com) &#x2F;item may contain (steve-yegge.blogspot.com) content
pluto_modadic大约 2 个月前
I&#x27;m going to guess this is satire unless sourcegraph only has &quot;junior&quot; developers, on &quot;junior&quot; pay.<p>The article tells me something unfortunate about the appropriateness about ever buying software from this person; based on how they write.
DanHulton大约 2 个月前
As always, citation needed.<p>(Also, grain of salt required, because this is a blatant marketing post.)<p>Look, I&#x27;ve been hearing &quot;the models will get better and make these core problems go away&quot; since it become common to talk about &quot;the models&quot; at all. Maybe they will some day! But also, and critically, maybe they won&#x27;t.<p>You also have to consider the future where some companies spend an additional $50-100k per developer and they DON&#x27;T see any of this supposed increase in performance, if these &quot;trust me, it&#x27;ll happen this time&quot; promises don&#x27;t come true. This is the kind of bet that can CRATER companies, so it&#x27;s not surprising to see some hesitation here, a desire to see if the football will be again yanked away.<p>Plus, and I believe most damningly, this article appears to be engaging in the classic technocratic failure mode: mistaking social problems for technical ones.<p>Obviously, yes, developers engage in solving technical problems, but that is not all they do, and at the higher level, that becomes the least of what they do. More and more, a good developer ensures that they are solving the RIGHT problem in the RIGHT WAY. They&#x27;re consulting with managers, (ideally) users, other teams, a whole host of people to ensure the right thing is built at the right time with the right features and that the right sacrifices are being made. LLMs are classically bad at this.<p>The author dismissively calls this &quot;getting stuck&quot;, and handwaves the importance of it away, saying that the engineer will be able to unstuck the model at first (which, if we&#x27;re putting armies of &quot;vibe coding&quot; junior engineers in charge of the LLMS, who&#x27;ve not had time enough in their career to develop this skill, HOW?), and then makes the classic claim &quot;but the models will get better&quot;, and predicts the models will eventually be able to do it (which, if this is an intractable problem with LLMS -- and so far evidence has been leaning this way -- again, HOW?).<p>Forgive that apalling grammar. I am het up. But note well what I&#x27;m doing: I&#x27;m asking &quot;should we even be doing this?&quot; Which is something these models a) will have to do well to accomplish what the author insinuates they will, and b) have been persistently terrible at.<p>I&#x27;m going to remain skeptical for now, since it seems that&#x27;s my one remaining superpower versus these LLMs, and I guess I&#x27;m going to need to keep that skill sharp if I want to avoid the breadline in this author&#x27;s future. =)
sfjailbird大约 2 个月前
LOL, pretty good, he had me going a few times. I&#x27;m sure a surprising amount of people will actually take this seriously, showing how ludicrous the situation is right now.
haburka大约 2 个月前
One of the funniest things I’ve read in a while. Also full of some truths. I think learning how to use AI will become a core part of being a dev but I seriously doubt they’ll have anywhere near the competency of solving a problem that a junior engineer has. They can certainly write code like one though.<p>I really recommend this to anyone reading - if you haven’t tried using cursor or copilot, check them out. It makes writing code less tedious.
lcnPylGDnU4H9OF大约 2 个月前
&gt; We think the flat hiring right now is just companies signalling that they don’t know what to do yet.<p>The whole article seems to be written disingenuously for the junior developer audience but this one kinda irked me: the flat hiring is because interest rates are high and has nothing to do with companies figuring out what to do with vibe coding.<p>On topic, nothing in this article suggests anything fundamentally useful about vibe coding other than it being an easier way to start for juniors and entry-levels. If you are a junior, go ahead and keep vibe coding but also do your best to understand the code you’re given. I strongly suspect that will (continue to) be something that makes people stand out.
DonHopkins大约 2 个月前
Is it just me, or has anyone else noticed that:<p>1) Cursor has been crashing several times an hour for me recently.<p>2) Cursor seems to ignore .cursorrules files. I&#x27;m using the json format that&#x27;s supposed to let you filter on file name patterns (although how that works for cross-cutting agent stuff I don&#x27;t know).<p>3) Cursor is obsessed with making sketchy iffy defensive code checking for the most recent symptom and trying to guess and shart its way out of it instead of addressing the real problem? And it&#x27;s extremely hard to talk it out of doing that, I have to keep reminding it and admonishing it to cut it the fuck out, fail instead of mitigate, address the root cause not the symptoms, and stop trying to close the barn door after all the horses have escaped. It&#x27;s as of it was only trained on Stack Overflow and PHP manual page discussions.
mwkaufma大约 2 个月前
Nothing smells like bullshit quite like a data-less &quot;graph&quot;