ckocagil

I prefer installing a generic fan that's one size bigger but lower RPM instead of a fancy Noctua. You end up with similar or higher CFM. Downside is it may require drilling extra holes. But it's much cheaper and I didn't measure but I bet it's quieter too.

And their most impactful application will be training smaller models for military drones to detect, target and kill humans. sigh...

Every time I'm about to be excited about AI the cynic in me slaps me back to reality. I imagine the endless possibilities of AI for disabled people, then realize that it's going to leave millions of people disabled instead.

Why fix your appliances and infrastructure and give you healthcare when the plan is to obsolete the whole lot of you?

The tangible hardware difference is the MMU. S3 didn't have a proper one. This new chip does. But the biggest difference is the intent. Microcontrollers focus on peripherals and are used for interfacing, data conversion and signal processing. They're meant to run either bare metal firmware or small RTOS. OTOH microprocessors have in addition to an MMU big caches, high IPC numbers, wide bus lanes, everything you need to run different workloads as fast as possible.

Technically a chip can excel at both.

They're not single board computers. They're microcontrollers. Or rather they were until this new chip was introduced.

Originally it was the more capable successor to ESP8266, the super cheap chip containing a high speed core with wifi and an IP stack. At one point it powered a huge chunk of all smart devices on the market. Then the original ESP32 was introduced as the successor, with BLE and two cores which were even faster. That made it a very capable chip for tasks demanding compute. After the original ESP32 they decided to add variety and make a whole "ESP32 family" and yes there's quite a lot of variety these days. Even the ISA and cores changed.

Isn't that the point? The camera glasses add a 10x harassment factor.

I'd rather go the other way and entirely decouple my online identity from Google (or Apple/Meta/Microsoft). Much better than constantly worrying about some crappy AI deciding to ruin your life with no recourse.

Are you talking about router firmware or NIC drivers? There's a slim chance of vendors opening drivers, but firmware is never going to happen.

>The stated goals of Wi-Fi 8 include a 25% increase in throughput at different signal-to-interference-and-noise ratio (SINR) levels, reduce latency by 25% for the 95th percentile scenarios with latency

So it is getting faster (in both senses) and doing so in scenarios that actually matter.

Because it's notoriously hard to benchmark LLMs. Ultimately every benchmark is different and measures different things. This is why companies that make LLM models have private benchmarks - they find the areas where the model is weak and make that their goal.

Isn't Q8 way overkill these days? I see many graphs showing Q4 or Q5 having less than %1 deviation. Nvidia's NVFP4 Qwen quantization should be even better due to its better training methods.

aka a stroboscopic measurement,

but I don't think it will work well for this case.

Late by a decade or more (JSR310 was released in 2014), but still a good development. I've tried convincing colleagues to use js-joda in the past but they thought they were keeping it simple by sticking to moment.js. They weren't.

Isn't the US stock market betting on AGI and superintelligence in the next decade? Maybe a lower population won't be as big of an issue, or even an advantage.