frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Jim Fan calls pixels the ultimate motor controller

https://robotsandstartups.substack.com/p/humanoids-platform-urdf-kitchen-nvidias
1•robotlaunch•2m ago•0 comments

Exploring a Modern SMTPE 2110 Broadcast Truck with My Dad

https://www.jeffgeerling.com/blog/2026/exploring-a-modern-smpte-2110-broadcast-truck-with-my-dad/
1•HotGarbage•2m ago•0 comments

AI UX Playground: Real-world examples of AI interaction design

https://www.aiuxplayground.com/
1•javiercr•3m ago•0 comments

The Field Guide to Design Futures

https://designfutures.guide/
1•andyjohnson0•3m ago•0 comments

The Other Leverage in Software and AI

https://tomtunguz.com/the-other-leverage-in-software-and-ai/
1•gmays•5m ago•0 comments

AUR malware scanner written in Rust

https://github.com/Sohimaster/traur
3•sohimaster•7m ago•0 comments

Free FFmpeg API [video]

https://www.youtube.com/watch?v=6RAuSVa4MLI
2•harshalone•7m ago•1 comments

Are AI agents ready for the workplace? A new benchmark raises doubts

https://techcrunch.com/2026/01/22/are-ai-agents-ready-for-the-workplace-a-new-benchmark-raises-do...
2•PaulHoule•12m ago•0 comments

Show HN: AI Watermark and Stego Scanner

https://ulrischa.github.io/AIWatermarkDetector/
1•ulrischa•13m ago•0 comments

Clarity vs. complexity: the invisible work of subtraction

https://www.alexscamp.com/p/clarity-vs-complexity-the-invisible
1•dovhyi•14m ago•0 comments

Solid-State Freezer Needs No Refrigerants

https://spectrum.ieee.org/subzero-elastocaloric-cooling
1•Brajeshwar•14m ago•0 comments

Ask HN: Will LLMs/AI Decrease Human Intelligence and Make Expertise a Commodity?

1•mc-0•16m ago•1 comments

From Zero to Hero: A Brief Introduction to Spring Boot

https://jcob-sikorski.github.io/me/writing/from-zero-to-hello-world-spring-boot
1•jcob_sikorski•16m ago•1 comments

NSA detected phone call between foreign intelligence and person close to Trump

https://www.theguardian.com/us-news/2026/feb/07/nsa-foreign-intelligence-trump-whistleblower
7•c420•16m ago•1 comments

How to Fake a Robotics Result

https://itcanthink.substack.com/p/how-to-fake-a-robotics-result
1•ai_critic•17m ago•0 comments

It's time for the world to boycott the US

https://www.aljazeera.com/opinions/2026/2/5/its-time-for-the-world-to-boycott-the-us
3•HotGarbage•17m ago•0 comments

Show HN: Semantic Search for terminal commands in the Browser (No Back end)

https://jslambda.github.io/tldr-vsearch/
1•jslambda•17m ago•1 comments

The AI CEO Experiment

https://yukicapital.com/blog/the-ai-ceo-experiment/
2•romainsimon•19m ago•0 comments

Speed up responses with fast mode

https://code.claude.com/docs/en/fast-mode
3•surprisetalk•22m ago•0 comments

MS-DOS game copy protection and cracks

https://www.dosdays.co.uk/topics/game_cracks.php
3•TheCraiggers•23m ago•0 comments

Updates on GNU/Hurd progress [video]

https://fosdem.org/2026/schedule/event/7FZXHF-updates_on_gnuhurd_progress_rump_drivers_64bit_smp_...
2•birdculture•24m ago•0 comments

Epstein took a photo of his 2015 dinner with Zuckerberg and Musk

https://xcancel.com/search?f=tweets&q=davenewworld_2%2Fstatus%2F2020128223850316274
12•doener•25m ago•2 comments

MyFlames: View MySQL execution plans as interactive FlameGraphs and BarCharts

https://github.com/vgrippa/myflames
1•tanelpoder•26m ago•0 comments

Show HN: LLM of Babel

https://clairefro.github.io/llm-of-babel/
1•marjipan200•26m ago•0 comments

A modern iperf3 alternative with a live TUI, multi-client server, QUIC support

https://github.com/lance0/xfr
3•tanelpoder•27m ago•0 comments

Famfamfam Silk icons – also with CSS spritesheet

https://github.com/legacy-icons/famfamfam-silk
1•thunderbong•28m ago•0 comments

Apple is the only Big Tech company whose capex declined last quarter

https://sherwood.news/tech/apple-is-the-only-big-tech-company-whose-capex-declined-last-quarter/
3•elsewhen•31m ago•0 comments

Reverse-Engineering Raiders of the Lost Ark for the Atari 2600

https://github.com/joshuanwalker/Raiders2600
2•todsacerdoti•32m ago•0 comments

Show HN: Deterministic NDJSON audit logs – v1.2 update (structural gaps)

https://github.com/yupme-bot/kernel-ndjson-proofs
1•Slaine•36m ago•0 comments

The Greater Copenhagen Region could be your friend's next career move

https://www.greatercphregion.com/friend-recruiter-program
2•mooreds•36m ago•0 comments
Open in hackernews

I Tested the M5 iPad Pro's Neural-Accelerated AI, and the Hype Is Real

https://www.macstories.net/stories/ipad-pro-m5-neural-benchmarks-mlx/
33•alwillis•2mo ago

Comments

bigyabai•2mo ago
Like prior iPad GPU upgrades, most customers will never notice until it reaches the Mac lineup. Another victim of the iOS/iPadOS capability bottleneck.
joakleaf•2mo ago
Related and test on MacBook Pro M5 vs M4:

https://machinelearning.apple.com/research/exploring-llms-ml...

"Exploring LLMs with MLX and the Neural Accelerators in the M5 GPU"

seanmcdirmid•2mo ago
Note this is the M5, not even the M5 Pro and definitely not the M5 Max or M5 Ultra. If they are getting these improvements on low end M series, I’m sort of interested in what happens with the M5 max when it’s ready (I. Not holding out hope that the M5 ultra will be done anytime soon).
greatgib•2mo ago
Prompt ingested (time to first token) in "18 seconds" with the new chip... end of the joke.
bigyabai•2mo ago
At this rate, Apple will be primed to replace OpenAI for realtime use somewhere around 2045.
seanmcdirmid•2mo ago
That is actually a reasonable prediction. By 2045, even tablets and high end phones will be able to locally run large enough models for real time chat.
seec•2mo ago
But don't forget, "the hype is real". Apple fanboys are something else.

At this point it's not even clear that the LLMs will have much use outside of chatbots spitting questionable reality faking but it's pretty clear that the small local models are largely useless. Outside of the poor user experience the lack of large dataset make them an exercice out technical feasibility more and anything else.

Apple feeling like they have to partner with Google to replace their own "Apple Intelligence" should tell you everything you need to know about local AI but I guess believers gotta believe.

alwillis•2mo ago
> Apple feeling like they have to partner with Google to replace their own "Apple Intelligence"

They’re not replacing Apple Intelligence; the partnership with Google is for the backend of Siri.

seec•2mo ago
Yeah but this is what I mean. They are unable to produce a decent AI solution so they have to outsource. And if it was just about being limited by on device processing power, they could just switch to using their cloud infrastructure.

To me that's just dumb. They are spending money for a competitor solution just to keep control and appearances. At this point they could just forget about AI altogether and/or open their OSs to let others integrate their solutions. But Apple is way too proud to admit being wrong and we will get a half-assed solution that will both cost them money and be inferior to the competition because of their own self-imposed limitations.

They have been riding on their privacy/security narrative for a while, yet they work with governments and ads/tracking hasn't gone anywhere. And they'll have to open up the App Store golden goose at some point anyway. They should recognise when the strategy isn't working for the long run and switch it up...

But they have tried to compete with Google for maps, it is still largely inferior in many ways and they seem to be fine with that. So I think it will end up in a similar situation.

alwillis•2mo ago
You know Google pays Apple ~$20 billion a year to be the default search engine, right? The rumored deal with Google is $1 billion, so Apple is still up $19 billion.

There's only three frontier model makers: OpenAI, Anthropic and Google. So everyone that's using a frontier model behind the scenes is using one of these three.

It's likely whatever happens, it'll be running on Apple's Private Compute Cloud [1]—it won't be mined for ads, etc.

What you don't understand is on-device AI isn't going away due to using Gemini for Siri; they're two separate products. Using Gemini is the fastest way to add "world knowledge" to Siri.

[1]: https://security.apple.com/blog/private-cloud-compute/

seec•2mo ago
I do understand, and I mostly agree with your perspective. But narrowing the problem to money isn't really helpful. Sure, Apple will be able to make a shitton of money regardless.

For consumers, it would be more beneficial if Apple could actually use all the money they get to really build standalone products. It feels like Apple is not trying very hard (or has become too incompetent) to make a viable competitive product because they profit enough regardless. Don't you think it would be better if Apple were able to make a search engine against Google instead of just pocketing the profit?

I'm not saying that on-device AI is going away because of Gemini. I just mean that Apple's strategy wasn't good, and they wasted resources on a dead end (a common trend with Apple recently), and they will rely on a competitor until they can provide their own in-house solution (that may or may not be good). They marketed the shit out of their Apple Intelligence for their hardware, and that ended up not being good.

The Private Cloud Compute push does not inspire great confidence because it basically relies on their hardware, just at a bigger scale. If they can't get good results with a Mac Studio, it's not clear the results will be that much better. It seems like the approach has a built-in limitation from both a hardware and software standpoint. Apple hardware is only decent at inference, and that's only when going for memory-heavy models and comparing to single consumer GPU machines. I think they have to get the model from Google in part because even if they have in-house talent, they would need to rely on 3rd-party hardware anyway.

They are not very competitive with local AI for average consumer devices, and they are not competitive with local AI at the high end (prompt processing). You can load high-quality models if you spend lots of money on their unified RAM, but it's slow. The Nvidia DGX, as well as the newer APUs from AMD, are rendering their memory advantage somewhat irrelevant for the midrange. And they can't provide competitive cloud AI for consumers who will never pay for such powerful hardware. So why should we pay more money for Apple hardware when you are going to end up using their competitors technology ? It's like buying an item from a white-label brand on Amazon that is just rebadging generic Chinese stuff. There is nothing wrong per se, but you are just giving money to a useless middleman; it's not efficient, and in a way, you get scammed.

You mention privacy as if it were still relevant. It's all moral posturing and virtue signaling from Apple. As I said, they don't work any less with governments than the other tech company, which is the real privacy you want. As for ads/tracking, not only does using their hardware not at all prevent that, but they are starting to sell ads as well, just in Maps for now, but it's likely to spread.

This is why I talked about Maps. Apple does not offer a competitive solution; the routing is pretty good, but today we use mapping apps for discovery and reviews, and personally, I also use transit and bicycle routes. Apple Maps is really not a decent equivalent, and I say this as someone who had a lot of hope when Apple Maps came out and used it almost exclusively for as long as I could. Even using both a Mac and an iPhone, I find the Google solution more convenient and useful. I live in the EU, where Google was forced to remove Maps links in their search engine, which was quite convenient. But the Maps macOS app is even more useless, so it might as well not exist. Poor data, poor UI, poor performance, it's just plain bad.

So in the end I use Google Maps, and the privacy argument is largely pointless because Google gets my data anyway. This is the same thing for needing to use Google Search, YouTube, and various other Google properties. Ultimately, Apple receives large sums of money for nothing of value, and they continue to act morally as though they were providing their customers with satisfactory service. It's effortless to paint competitors negatively, but if you can't offer an adequate solution, then nobody will benefit except yourself. And they have the money, so it's either greed or incompetence.

And this is why I said that about their AI marketing. It's largely useless compared to cloud AI that most consumers get to use on competitors devices/solutions. Even for strictly local AI, they are not very competitive at any price point. Best-case scenario, it's a built-in solution that could be decent if they figure out the model part correctly.

The reality is that the hype is widely overblown, but I have been buying Apple hardware since the early 2000s, so I'm used to it. But they are just losing relevancy, and at some point having good chips and nice premium hardware won't be enough.

alwillis•2mo ago
> Prompt ingested (time to first token) in "18 seconds" with the new chip... end of the joke

18 seconds on the M5; 4.4x times faster than the previous M4 running one of Qwen’s 8 billion parameter local models.

That’s quite impressive for a tablet and faster than most laptops available today.