frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

Open in hackernews

AI Efficiency? Give Me a Break

https://luolink.substack.com/p/ai-efficiency-give-me-a-break
12•luolink•1d ago

Comments

luolink•1d ago
I recently watched an AI expert on stage enthusiastically proclaiming that "AI will liberate human productivity" to thunderous applause. Sitting at my desk, staring at the 37th "Essential AI Tools You Must Learn" popup of the day, I couldn't help but think: Give me a break. I'm ten times more exhausted now than I ever was before.

An Endless Stream of New Tools and Sleepless Anxiety Nights I remember when I first encountered AI tools. When ChatGPT burst onto the scene, I was as excited as an explorer discovering a new continent. I spent an entire weekend mastering prompt engineering techniques, convinced I was finally riding the wave of the future.

And then what happened?

A month later, GPT-4 launched. All those prompt strategies I'd painstakingly learned suddenly felt outdated.

Two months later, Claude appeared, supposedly superior for writing tasks.

Three months later, Midjourney updated to V6, rendering all my carefully memorized parameters obsolete.

Four months later, domestic large language models sprouted like mushrooms after rain — Wenxin Yiyan, Tongyi Qianwen, iFlytek Spark... each claiming to "better understand Chinese."

My bookmarks folder now houses 128 AI tool websites. From writing to image generation, coding to video creation, data analysis to presentation design. Every single one has been shared by someone on social media with the caption "Don't learn this and you'll fall behind!"

The most maddening part? These tools update faster than I change my underwear. Just when I've familiarized myself with an interface, it gets redesigned overnight. Just when I've memorized a workflow, next week brings "revolutionary new features."

My morning routine no longer starts with coffee

smartmic•1d ago
I wrote about my reaction to the AI flood yesterday[0]. In short: I'm not participating, but waiting to see what happens. Your abilities will remain intact, so make sure that your mind remains intact as well.

[0] https://news.ycombinator.com/item?id=44892946

HPsquared•1d ago
Sounds like overfitting. Ironic!
sirwhinesalot•1d ago
> All those prompt strategies I'd painstakingly learned suddenly felt outdated.

This is the critical bit, the bit that makes it impossible to "fall behind". There's no actual useful knowledge you can learn right now that'll stay relevant in the future, so don't bother!

Just use these things naively. Investment on your end on how to "use them properly" should be close to 0. If they produce useful output with the most naive prompt imaginable, wonderful. If they don't, do something else.

All those people running an army of agents on their codebase with carefully curated rules file and whatever else? They're vibing their way off a cliff. No need to worry about them. You'll probably get to the same place they do by just using the autocomplete mode in VSCode with a braindead cheap model you get for free.

wat10000•1d ago
Ignore them. I’ve played around with this stuff so I have some idea of what it’s all about, but only minimally. I haven’t used it for work at all. Despite this, I remain gainfully employed, and I’m convinced that I’m still going to be fielding offers at the age of 90 because pretty much none of the programmers diving into this stuff will have any idea how a computer actually works. You can keep your toes in the water without much commitment.
giik•1d ago
Amen!
Havoc•1d ago
> My phone pushes at least 20 of these "anxiety generators" daily.

Unsubscribe / opt out / uninstall?

If you set yourself up to be bombarded then that’s exactly what happens

lentil_soup•1d ago
I don't understand the rush to adopt things if they're not really giving you a boost right now, some people still use vi and hammers, both invented by early humans. Adopting a new tool while a new one comes out every week is fishing in a troubled river. It's perfectly fine to wait until things settle and then learn whatever is actually left, I hightly doubt you're going to lose your job because you didn't learn something 2 days ago.
dinfinity•1d ago
I would say:

1. Spend time on the fundamentally new things through whatever tool provides it.

2. Always try to prevent 'vendor' lock-in. Think about portability and reusability.

A lot of the stuff that now "works well" when working with AI assistants and tools is stuff that was always a good idea and always worked well. Write Once, Read Many; Provide good specification; Communicate clearly; Automate repeated tasks; etc.

If you update your workflow in a fundamental manner and don't jump from investing 100% in tool X to 100% in tool Y redoing a lot of shit, you improve it efficiently.

mquander•1d ago
Maybe if you are tired of chasing AI hype you could start by not literally AI-generating big clickbait articles about AI for social media?
ApeWithCompiler•1d ago
I agree to this oppinion. But I furthermore want to add: It's not just about unsubscribing or ignoring the noise. In fact, unsubscribing or ignoring means to break with the premise sold, to begin with. If it is true, that one must learn this or that to stay employed, he can not just "unsubscribe" from it.

Additionally the notion of productivity in our industry is problematic. While working with machines, somehow we developed their standard of productivity as our ideal. May it be the pressure of competition for companies and employees alike, but the current notion is not sustainable. Exaggerated, but what I think: Turn away from two week sprints and work in a quaterly waterfall. Give developers a break, a constant plan and environment.

athrowaway3z•1d ago
> "Master This AI Prompt Template and Earn $10K Monthly!" "I Made Enough for My First House in 3 Months of AI Side Hustles!" "Former Google Engineer Reveals How to Surf the AI Wave!"

I once had the displeasure of talking with a guy who essentially sold FOMO "get rich" schemes. What he sold was expensive and imo useless - but mostly suckers signed up, so the inevitable 1/5 reviews were the exception.

He made a decent living with it, but in terms of hours worked it really wasn't a good deal.

But one of the things he was self-aware enough to laugh about, was that in literal years of giving his course to hundreds of people, only once had somebody asked him: "The marketing claims you're rich and successful so why are you still giving this course?"

If you read those headlines and find a tingle of FOMO coming on, ask yourself that question.

Or more general: When would you personally ever write a blog designed to triggers FOMO, and does that match the results promised in the title?

nirui•1d ago
Interestingly, or rather disappointingly, it looked like the "Fear-Mongering Marketing" is actually working.

YouTube creator ColdFusion recently uploaded a video titled "Gen Z Graduates Are in Crisis" (https://www.youtube.com/watch?v=yVc2ZhECTMg), which is about the impact of AI against Gen Zs. In it he talked about how AI is worsening job market.

I'm thinking, since the AI marketers are hyping up AI tools, maybe it has created an illusion that worker might not be needed in the near future, then it propagated outward like that. Considering how sluggish the job market already was, it's not surprising things could went on like this.

My guess is, it will continue the trend, until the moment the bubble pops, if the bubble is not just a joke.

But then I looked at how many people are paying for AI tools right now...

HR paying AI tools to read resume, job seeker paying AI tools to write resume; teacher paying AI tools to review homeworks, student paying AI tools to write homework; Academic journals paying AI tools to scan for cheats, paper factories paying AI tools to "optimize" their "paper"... Oh, and people vibe coding their project on GitHub, etc etc.

Well, at least for now it's fun to watch how things are panning out. Give yourself some laughter before the pain comes.

PuTTY has a new website

https://putty.software/
100•GalaxySnail•2h ago•59 comments

Best Practices for Building Agentic AI Systems

https://userjot.com/blog/best-practices-building-agentic-ai-systems
84•vinhnx•3h ago•39 comments

The future of large files in Git is Git

https://tylercipriani.com/blog/2025/08/15/git-lfs/
292•thcipriani•10h ago•117 comments

AI is different

https://www.antirez.com/news/155
175•grep_it•2d ago•166 comments

I accidentally became PureGym’s unofficial Apple Wallet developer

https://drobinin.com/posts/how-i-accidentally-became-puregyms-unofficial-apple-wallet-developer/
258•valzevul•19h ago•53 comments

Show HN: Edka – Kubernetes clusters on your own Hetzner account

https://edka.io
298•camil•12h ago•93 comments

Occult books digitized and put online by Amsterdam’s Ritman Library

https://www.openculture.com/2025/08/2178-occult-books-now-digitized-put-online.html
367•Anon84•14h ago•142 comments

Do Things That Don't Scale (2013)

https://paulgraham.com/ds.html
297•bschne•15h ago•91 comments

Deep-Sea Desalination Pulls Fresh Water from the Depths

https://www.scientificamerican.com/article/deep-sea-desalination-pulls-drinking-water-from-the-depths/
24•noleary•3d ago•2 comments

OpenBSD is so fast, I had to modify the program slightly to measure itself

https://flak.tedunangst.com/post/is-OpenBSD-10x-faster-than-Linux
150•Bogdanp•11h ago•111 comments

ADHD drug treatment and risk of negative events and outcomes

https://www.bmj.com/content/390/bmj-2024-083658
214•bookofjoe•15h ago•260 comments

Prompting by Activation Maximization

https://joecooper.me/blog/activation/
4•thatjoeoverthr•2d ago•0 comments

Launch HN: Embedder (YC S25) – Claude code for embedded software

85•bobwei1•12h ago•38 comments

The electric fence stopped working years ago

https://soonly.com/electric-fences/
195•stroz•14h ago•84 comments

Porting Gigabyte MZ33-AR1 Server Board with AMD Turin CPU to Coreboot

https://blog.3mdeb.com/2025/2025-08-07-gigabyte_mz33_ar1_part1/
61•pietrushnic•9h ago•0 comments

TextKit 2 – The Promised Land

https://blog.krzyzanowskim.com/2025/08/14/textkit-2-the-promised-land/
63•nickmain•8h ago•21 comments

Show HN: Prime Number Grid Visualizer

https://enda.sh/primegrid/
80•dduplex•2d ago•38 comments

A privacy VPN you can verify

https://vp.net/l/en-US/blog/Don%27t-Trust-Verify
67•MagicalTux•7h ago•59 comments

Model intelligence is no longer the constraint for automation

https://latentintent.substack.com/p/model-intelligence-is-no-longer-the
49•drivian•15h ago•56 comments

Recto – A Truly 2D Language

https://masatohagiwara.net/recto.html
102•mhagiwara•3d ago•49 comments

ARM adds neural accelerators to GPUs

https://newsroom.arm.com/news/arm-announces-arm-neural-technology
147•dagmx•3d ago•29 comments

Bullfrog in the Dungeon

https://www.filfre.net/2025/08/bullfrog-in-the-dungeon/
110•doppp•13h ago•26 comments

Compiler Bug Causes Compiler Bug: How a 12-Year-Old G++ Bug Took Down Solidity

https://osec.io/blog/2025-08-11-compiler-bug-causes-compiler-bug/
124•luu•4d ago•62 comments

A mind–reading brain implant that comes with password protection

https://www.nature.com/articles/d41586-025-02589-5
34•gnabgib•1d ago•18 comments

Claude Opus 4 and 4.1 can now end a rare subset of conversations

https://www.anthropic.com/research/end-subset-conversations
207•virgildotcodes•9h ago•292 comments

Vaultwarden commit introduces SSO using OpenID Connect

https://github.com/dani-garcia/vaultwarden/pull/3899
160•speckx•17h ago•85 comments

America's Stock-Market Dominance Is an Emergency for Europe

https://www.wsj.com/finance/investing/americas-stock-market-dominance-is-an-emergency-for-europe-627b3221
12•mudil•1h ago•1 comments

Open hardware desktop 3D printing is dead?

https://www.josefprusa.com/articles/open-hardware-in-3d-printing-is-dead/
645•rcarmo•17h ago•393 comments

Is air travel getting worse?

https://www.maximum-progress.com/p/is-air-travel-getting-worse
126•mhb•17h ago•209 comments

Thai Air Force seals deal for Swedish Gripen jets

https://www.scmp.com/news/asia/southeast-asia/article/3320828/us-f-16s-lose-out-thai-air-force-seals-us600-million-deal-swedish-gripen-jets
146•belter•8h ago•191 comments