Headworx

Managing energy consumption in battery-powered electronic gadgets is one of the most difficult engineering tasks. And no one has dine it 100% right. Low power consumption is only achieved through managing the device activity. A device may really be designed to and perform consuming manageable amount of energy, until once in a while it runs away and dies. A 2018 MacBook Pro runs for 8 hours under macOS and only 90 minutes under Windows. The reason? The accelerated graphic card (the ATI Radeon GPU) is tightly managed by macOS (by switching it off and using the integrated low-performance / low-power Intel GPU). Wen booting Windows, the Apple BIOS disables the Intel GPU so Windows has nothing to switch back to and consequently draws batteries at full speed. My Garmin Delta watch has a habit of draining its battery to death in the night when it repeatedly attempts to connect to the iPhone to sync data. I have a habit of leaving the watch on a kitchen table and taking the phone with me to t...

I should have known this time. This happened to me in the past multiple times. I bought an SSD drive which was 256GB on the outside and only 40GB inside. I bought suspiciously inexpensive 1TB microSD card which was way less than that in practice. I bought branded fragrances which were fake. And this time I bought AirPods Pro 2 for a too-good price. They were fake. Fake, but really hard to distinguish from originals. I just could not believe what I saw.  The AirPods looked and felt like the originals. Perhaps the audio quality was not that great, but that is always a bit subjective. Anyway I wanted them for my Mom. She has some minor hearing loss. I read all the recommendations about the Pro 2s stepping in the role of a hearing aid (with a firmware upgrade). Of course I wanted to try this feature, so all I was looking for right after pairing with my iPhone was the firmware update. The AirPods paired like the originals. Displayed all the native features, including the serial number,...

Musk's acquisition of Twitter started the first wave of exodus (or simply the Xodus) but it was not definitive. People were trying Mastodon or Facebook's Threads , but none of that reached the critical mass. It may be different this time, as the US presidential elections has polarized the audience to the levels previously unseen. I personally tried all those options but also sticked with Twitter. But over the last couple of months it has become clear the platform has clearly become algorithmically  biased and flooded with bots. Nothing in common with the officially claimed "free speech". Even to a casual reader like myself it has become mostly a sewage, in which even funny orange cat videos could not make it to the surface. The recent shift of users from x.com to Bluesky might have been sparked by Taylor Swift fans , but then it has accelerated like an avalanche. The Guardian was next . And now it seems like everybody is moving. Including some old friends of mine w...

Santa has visited me early this year. Something a bit unexpected but I love the gadget a lot: the iFixit portable soldering station . Battery (or power bank) powered soldering iron has been in my toolkit for some time. I had been using the TS80P . Which had been good for small field jobs but not really capable to fully replace a professional desktop station. The iFixit station is the best of the both worlds. Designed to sit nicely on a desk, with a specially designed station / controller / power bank (100W output, 2 output ports). And equally handy to take outside / to a car / or wherever there is a need to do some soldering and dragging a power cable would not be easy. They don't say it, but clearly the base has been designed with some additional tools available in future - as it has two independent USB-C power ports. The quality is top-notch: great ergonomic design and top grade materials like the super elastic cable. I'd say I'd recommend this as a Christmas present, but...

Technology capabilities have for years been outpacing real life adoption. The gap seems to be widening as on the technology front we are talking now about Artificial General Intelligence (AGI) while simple everyday things just don't work. Last week the airline lost my luggage and the recovery started nicely as I got a personalized message indicating they had already found it and it was n its way. The message prompted me to a simple web form to enter my temporary address they should deliver the bag to. This was all on Oct 28. i was staying at the hotel for 2 nights, so I put Oct 30 in the form. Early morning on the next day (Oct 29) I got a notification asking about the baggage delivery, as I was supposedly "about to leave the temporary address". Surprised, I went back to the form which was indeed showing Oct 29. I was sure I entered Oct 30... But OK, I did this again. Picked Oct 30 from a nicely crafted date picker. The form was still showing Oct 29 The regular tricks suc...

Bitcoin, and particularly the blockchain technology was supposed to change the world. That was 15 years ago. And the world went crazy indeed. Particularly in the startup / VC world blockchain was the only thing that counted. Fast forward to 2024 and we barely have a trace of that. I mean Bitcoin still exists, mostly as a (very niche) speculative instrument. And also serves very well the dark communities. I complained on that back in 2021 . About half of Bitcoin transactions serve illegal activity. And even if legal - the speculative and mining for money part - it does not do any good either. Investment groups have been reigniting decommissioned power plants. Generating air pollution. And noise - to the extent that people decide to sue Bitcoin "mines" operators . Generally speaking - it is useless. Enter now the 2024 and the Artificial Intelligence craze. Myriads of promises and tons of "cool" demos. Again the startup and VCs have gone crazy about it. But the net ga...

    On October 7, 2024, Apple published a research paper that discusses the limitations of mathematical Reasoning in large language models (LLMs). This is yet another proof that LLMs are about language and have very little in common with understanding the underlying meaning. This is by the way quite similar to how human brains work, as we have (at least) two separate neural networks (and probably more) - one responsible for parsing and constructing language statements, and a completely separate one where the actual "thinking" happens. I posted a blog on that back in May. Long story short, the Apple researchers point out that despite the overall heralded progress in LLMs, they are still bound by limits of pattern matching, when they "convert statements to operations without truly understanding their meaning". This "suggests deeper issues in reasoning processes" that can't be helped with fine-tuning or other refinements. In other words there is still t...