frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Raspberry Pi's New AI Hat Adds 8GB of RAM for Local LLMs

https://www.jeffgeerling.com/blog/2026/raspberry-pi-ai-hat-2/
43•ingve•1h ago

Comments

dwedge•1h ago
> In practice, it's not as amazing as it sounds.

8GB RAM for AI on a Pi sounds underwhelming even from the headline

agent013•1h ago
A good illustration of how “can run LLM” ≠ “makes sense to run LLM”. A prime example of how numbers in specs don’t translate into real UX.
phito•1h ago
Sounds like some PM just wanted to shove AI marketing where it doesn't make sense.
buran77•1h ago
I think Raspberry lost the magic of the older Pis, they lost that sense of purpose. They basically created a niche with the first Pis, now they're just jumping into segments that others created and are already filled to the brim with perhaps even more qualified competition.

Are they seeing a worthwhile niche for the tinkerers (or businesses?) who want to run local LLMs with middling performance but still need full set of GPIOs in a small package? Maybe. But maybe this is just Raspberry jumping on the bandwagon.

I don't blame them for looking to expand into new segments, the business needs to survive. But these efforts just look a bit aimless to me. I "blame" them for not having another "Raspberry Pi moment".

ulnarkressty•1h ago
That niche was long taken over by cheap Chinese SBCs, so they have to innovate somehow. Their only advantage that remains is the community.
joe_mamba•58m ago
Yep. RPi foundation lost the plot a long time ago. The original RPi was in a league of its own when it launched since nothing like it existed and it was cheap.

But now if I want some low power linux PC replacement with display output, for the price of the latest RPi 5, I can buy on the used market a ~2018 laptop with a 15W quad core CPU, 8GB RAM, 256 NVME and 1080p IPS display, that's orders of magnitude more capable. And if I want a battery powered embedded ARM device for GPIO over WIFI, I can get an ESP32 clone, that's orders of magnitude cheaper.

Now RPi at sticker price is only good for commercial users since it's still cheaper than the dedicated industrial embedded boards, which I think is the new market the RPI company caters to. I haven't seen any embedded product company that doesn't incorporate RPis in its products now, or at least in their lab/dev/testing stage, so if you can sell your entire production stock to industrial users why bother making less money selling to consumers?

haunter•47m ago
Not just laptops but the used enterprise micro PCs from Dell, HP, and Lenovo. All the same small form factor with very low TDP You can have up to 32 or 64 GB RAMs depending on the CPU, dual or even triple disks if you want a NAS etc.
joe_mamba•40m ago
yeah, depends on what the used market looks like where you live. Here I see way more laptops for sale for cheap than those enterprise thin clients.

And the thin clients when they are for sale tend to have their SSDs ripped out by IT for data security, so then it's a hassle to go out and buy and extra SSD, compared to just buying a used laptop that already comes with display , keyboard, etc.

Mashimo•35m ago
I have seen quite a few the size of a Mac Studio / Intel Nuc, what are the device called that are the size of a pi?
qsera•37m ago
>I can buy on the used market a ~2018 laptop with a 15W quad core CPU, 8GB RAM, 256 NVME and 1080p IPS display, that's orders of magnitude more capable..

But it won't be as reliable, mostly motherboards won't last long.

joe_mamba•35m ago
Don't know what your source is for that, but that's not my experience, and i've had dozens of laptops through my hands due to my hobby.

The ticking timebomb lemons with reliability or design issues, will just die in the first 2-4 years like clockwork, but if they've already survived 6+ years without any faults, they'll most likely be reliable from then on as well.

ForHackernews•24m ago
Bathtub curve is extremely common https://en.wikipedia.org/wiki/Bathtub_curve
Mashimo•25m ago
I still use Pis in my 3d printers. Laptop would be too big, and a ESP could not run the software. "China clone" might work, but the nice part of the pi is the images available. It just works™

I'm also currently building a small device with 5" touchscreen that can control a midi fx padle of mine. It's just so easy to find images, code and documentation on how to use the GPIO pins.

Might be niche, but that is just what the Pi excels at. It's a board for tinkers and it works.

cyberax•23m ago
RPi will still have lower power consumption and is far more compact. And mechanically reliable.

I'm in the market to replace my aging Intel NUCs, but RPi is still cheaper.

TazeTSchnitzel•53m ago
The Raspberry Pi probably still has the advantage of an actually robust firmware/software ecosystem? The problem with SBCs has always been that the software situation is awful. That was the Raspberry Pi's real innovation: Raspbian and a commitment to openness.
vachina•26m ago
> The problem with SBCs has always been that the software situation is awful

Awful how? A SBC can take advantage of many software written from the dawn of x86.

crimsoneer•48m ago
Not everything needs to be for everyone. I think this is super cool - I run a local transcription tool on my laptop, and the idea of miniaturising it is super cool.
noodletheworld•36m ago
It is neat, and at 32GB it might be useful.

Almost nothing useful runs in 8.

This is the problem with this gen of “external AI boards” floating around. 8, 16, even 24 is not really enough to run much useful, and even then (ie. offloading to disk) they're so impractically slow.

Forget running a serious foundation model, or any kind of realtime thing.

The blunt reality is fast high memory GPU systems you actually need to self host are really really expensive.

These devices are more optics and dreams (“itd be great if…”) than practical hacker toys.

simjnd•1h ago
Ah yes, more AI e-waste
moffkalast•1h ago
> The Pi's built-in CPU trounces the Hailo 10H.

Case closed. And that's extremely slow to begin with, the Pi 5 only gets what, a 32 bit bus? Laughable performance for a purpose built ASIC that costs more than the Pi itself.

> In my testing, Hailo's hailo-rpi5-examples were not yet updated for this new HAT, and even if I specified the Hailo 10H manually, model files would not load

Laughable levels of support too.

As another datapoint, I've recently managed to get the 8L working natively on Ubuntu 24 with ROS, but only after significant shenanigans involving recompiling the kernel module and building their library for python 3.12 that Hailo for some reason does not provide outside 3.11. They only support the Pi OS (like anyone would use that in prod) and even that is very spotty. Like, why would you not target the most popular robotics distro for an AI accelerator? Who else is gonna buy these things exactly?

renewiltord•59m ago
What’s the current state of the art in low power wake word and speech to text? Has anyone written a blog post on this?

I was able to run a speech to text on my old Pixel 4 but it’s a bit flaky (the background process loses the audio device occasionally). I just want to take some wake word and then send everything to remote LLM and then get back text that I do TTS on.

geerlingguy•56m ago
Maybe not SOTA but the HA Voice Preview Edition [1] in tandem with a Pi 5 or some similar low-power host for the Piper / Whisper pipeline is pretty good. I don't use it but was able to get an Alexa/Google Home-like experience going with minimal effort.

I was only using it for local Home Assistant tasks, didn't try anything further like retrieving sports scores, managing TODO lists, or anything like that.

[1] https://www.home-assistant.io/voice-pe/

monocasa•55m ago
Wake word can be tiny. Like 10k weights and can run on an esp32 or similar with plenty of compute to spare.

TinyML is a book that goes through the process of building a wake word model for such constrained environments.

huntercaron•55m ago
Glad Jeff was critical here they need a bit of a wake up call it seems.
speedgoose•45m ago
Is there any usefulness with the small large language models, outside perhaps embeddings and learning?

I fail to see the use-case on a Pi. For learning you can have access to much better hardware for cheaper. Perhaps you can use it as a slow and expensive embedding machine, but why?

cmpxchg8b•35m ago
8GB? What is this, an LLM for ants?
kotaKat•18m ago
"For example, the Hailo 10H is advertised as being used for a Fujitsu demo of automatic shrink detection for a self-checkout."

... why though? CV in software is good enough for this application and we've already been doing it forever (see also: Everseen). Now we're just wasting silicon.

NamePhi – AI-Powered Domain Name Generator for Brandable Identities

https://www.namephi.com/
1•quillzhou•38s ago•0 comments

Reconstructability as a Threshold Question in AI-Mediated Representation

https://www.aivojournal.org/reconstructability-as-a-threshold-question-in-ai-mediated-representat...
1•businessmate•1m ago•1 comments

Open Source Strategies

https://tempestphp.com/blog/open-source-strategies
1•brentroose•3m ago•0 comments

ATT Room 641 [pdf]

https://jnslp.com/wp-content/uploads/2014/05/On-the-Bulk-Collection-of-Tangible-Things.pdf
1•rballpug•5m ago•1 comments

Two AI researchers are now funded by Solana

https://ghuntley.com/solana/
1•ghuntley•7m ago•0 comments

MultiLang‑ASM:A Universal Assembly Platform for Kids,Students,Kernel Developers

1•neuroosgenesis•9m ago•0 comments

Simpler than Photoshop but for free AI Landscaping

https://hadaa.pro/
2•Fh_•11m ago•1 comments

Photos Capture the Breathtaking Scale of China's Wind and Solar Buildout

https://e360.yale.edu/digest/china-renewable-photo-essay
3•mrtksn•11m ago•0 comments

Physics of foam resembles AI training

https://phys.org/news/2026-01-physics-foam-strangely-resembles-ai.html
2•sebg•12m ago•0 comments

Formbase – open‑source back end for HTML forms

https://github.com/formbase-dev/formbase
2•ephraimduncan•18m ago•0 comments

DB Branching in Mirrord

https://metalbear.com/blog/db-branching/
1•aviramha•19m ago•0 comments

A simple TUI video converter for the AV1 codec

https://github.com/framicheli/av1converter
1•francescomi•22m ago•0 comments

Show HN: Free and open sourced prompt skills game for kids

https://prompts.chat/kids
1•fka•23m ago•0 comments

Mission planning system will break before your rockets do

https://blog.satsearch.co/2026-01-14-spotlight-your-mission-planning-system-will-break-before-you...
1•kartikkumar•24m ago•0 comments

I built a tool to help me explain my work to clients automatically

https://delivrr.work/
1•ymfh•28m ago•1 comments

Show HN: 0xCal – A calorie tracker where you just describe what you ate

https://apps.apple.com/pl/app/0xcal-ai-calorie-tracker/id6749210009
1•namedix•28m ago•0 comments

Declarative YAML Workflow System for AI Agents

https://twitter.com/nedim0x01/status/2011733056709636471
1•nedim0x01•30m ago•1 comments

AI-powered automatic translation in WordPress (YouTube video tutorial)

https://gatoplugins.com/blog/youtube-video-tutorial-ai-powered-automatic-translation-in-wordpress
1•leoloso•33m ago•0 comments

LiteStar – a FastAPI rival (DTOs, controllers, msgpack)

https://litestar.dev/
1•begoon•33m ago•0 comments

Second Thoughts on James Burnham (George Orwell, 1946)

https://www.orwellfoundation.com/the-orwell-foundation/orwell/essays-and-other-works/second-thoug...
2•duggan•34m ago•0 comments

Semi-Automating 200 Pull Requests with Claude Code

https://blog.davisvaughan.com/posts/2026-01-09-claude-200-pull-requests/
1•sebg•35m ago•0 comments

The crisis whisperer: how Adam Tooze makes sense of our bewildering age

https://www.theguardian.com/business/2026/jan/15/the-crisis-whisperer-how-adam-tooze-makes-sense-...
1•6LLvveMx2koXfwn•37m ago•0 comments

Blacksmith – AI Powered Penetration Testing

https://github.com/yohannesgk/blacksmith
4•jchris280•42m ago•4 comments

Codeless Code – Fables and Koans for the Software Engineer

https://thecodelesscode.com/case/234
1•todsacerdoti•44m ago•0 comments

Napa: Powering Scalable Data Warehousing with Robust Query Performance at Google

https://research.google/pubs/napa-powering-scalable-data-warehousing-with-robust-query-performanc...
1•tosh•44m ago•0 comments

Bruce Perens: The Ham Radio Operator Who Made Trump Want Greenland

https://web.archive.org/web/20260115085443/https://www.linkedin.com/pulse/ham-radio-operator-who-...
1•weinzierl•45m ago•1 comments

Use Agents or Be Left Behind? A Personal Guide to Automating Your Own Work

https://timdettmers.com/2026/01/13/use-agents-or-be-left-behind/
4•sebg•46m ago•0 comments

Optimizing data throughput for Postgres snapshots with batch size auto-tuning

https://xata.io/blog/postgres-snapshots-with-batch-size-auto-tuning
1•gulcin_xata•47m ago•1 comments

Typical: TypeScript with type safety at runtime

https://typical.elliots.dev/
1•elliotshep•47m ago•0 comments

Me and the Machine

https://mitsuhiko.github.io/talks/me-and-the-machine/
2•obiefernandez•48m ago•0 comments