If you read the review you’d see they mention x64 emulation through things like GameNative. You can play lots of steam games on android devices now, the major limiting factor is the power of the hardware on android devices and driver support.
Yeah, for some reason I read high beams and just thought of the bright LEDs everyone has. Either way, I agree that it doesn’t make sense to blame high beams.
If those high beams are legal, then all that means is the robot can’t safely operate under normal/legal conditions. If it can’t operate safely, why is it allowed in public?
You can’t just claim “Availability Heuristic” without providing any other evidence.
Additionally, a lot of people’s issue with these things is they’re operating in a legal gray zone. They’re using up pedestrian space for companies benefits.
For Halloween I used an orangepi (cause of cost, but would not recommend due to support) to install an LCD eye in a skeleton that tracks people using a object detection model. Additionally it’s setup with some voice commands/interactions.
Overall that means holding 3+ ML models in ram (yolo, whisper, Vad, etc.). That basically uses up all the RAM I have and before optimizations was actually causing the chip to overheat (was able to fix that, but still hits the cpu pretty hard).
I’m not talking about price, I’m talking about generation.
I have solar panels, they are great. However, I get too much energy in the summer, not nearly enough in the winter, and no power at night. Since I’m on a grid that doesn’t matter too much for me, but highlights some of the issues.
Even if I installed a battery, I’d end up with extra power in the summer and not enough in the winter. Which just highlights the discrepancy in production. Batteries obviously help solve this, but some mixture of wind/nuclear/other is needed.
Solar still has some huge problems when it comes to seasonality and day/night cycle. You either need massive battery build out or some other form of energy to handle base loads and/or offset when solar isn’t producing enough.
Not saying solar isn’t part of the future, but we literally can’t just rely on solar.
I think it’s more likely they make it dangerous enough that large countries can’t afford to not have it. That’s what largely seems to be happening in the US and China, both countries probably see AI as both a weapon and defensive tool.
My response was only in regard to the speed in which something like IoT devices can operate.
You’re right that goals/directions of LLMs can be persisted anywhere, but a swarm or fleet of AI running on IoT devices will be severly hampered by compute (based on current models). Which was the main point I was making, in that there is only so much you can do to distribute their processing. Meaning you can parallelize the number of models, but can’t speed them up beyond a certain limit. This is based on the fact that some form of a layer and it’s weights must exist within the memory of a single device.
The downside to this limitation is that it is temporary, eventually IoT devices will probably be powerful enough, or the worms could just attack phones/laptops and easily get the compute they need. I think the real risk of IoT devices is low security mixed with low “supervision” (no one would easily notice their IoT device is doing extra processing).
This is like the like the “9 women in one month to make a baby” comment.
There is only so much you can distribute the layers of a model due to interconnectedness. And you can’t run multiple inferences at the same time because the next token relies on the previous.
You could be doing multiple things at once, but each task will run “100,000x” if the OC did their math right.
Glad to hear I’m not the only person who has had some difficulties. And yeah, you are right that it should be fine for hosting services, which is why I mentioned my specific use cases.
However, I think calling out lack of support is always a good thing as you never know what support you may need down the road. I definitely have run into some issues due to their supported OS I’m running and its surprisingly out of date for a board that only released a couple years ago.
I was going to complain that the orangepi wasn’t $350+, but didn’t realize the fully maxed 6 is that much. That’s crazy how much single board computers have gotten.
Just shared my experience on the post you replied to, but basically the support/documentation for the orangepi isn’t great and has cost me a lot of time and forced me to use workarounds to get some things working (GPIO, NPU, cameras, etc.)
OrangePi support is terrible. Maybe it’s better with the 5, but the 4 I’ve struggled with getting a library to work with the GPIO pins, documentation on the NPU is limited, and I can’t get it to recognize a standard serial camera.
It’s possible I’m just trying to do weird things, but it’s a much worse experience than I’ve ever had with a raspberry pi and I regret not spending more to just get the pi. I’ve lost probably 10+ hours just trying to deal with some things I’d consider basic, and had to create weird workarounds that hurt performance in other areas.
Prediction markets are just “sparkling” insider trading.
I didn’t think LibreOffice supported the latest versions of office formats though, so wouldnt it potentially break existing files if used?
Data centers need to run 24/7, so you aren’t able to meet the needs with just solar. Additionally Solar is very seasonal (more power during the summer).
Not saying there aren’t ways to making it work, but one natural gas plant vs a mixture of systems is probably cheaper and easier. Especially since the US is currently built around natural gas power.
If you like elder scrolls and dark fantasy you should try Tainted Grail Fall of Avalon
Pretty sure most countries do this now, sadly.
You do realize for many people (in the world), their phone is their only computer/access to the internet?