Think of a new born baby. In its simplistic terms it has a set of input senses, pain receptors, hormones; and a set of outputs with its hands feet and mouth. All these inputs are basically prioritized into a pain/pleasure level. Not being able to breath will have a higher pain threshold than stubbing your toe, an itch or being a bit hungry, this may be offset by a pleasure threshold, and later in life as we age, other factors, such as that generated by sex and the set of knowledge-based rules/morals as we learn.

The first LLM’s were basically single pass, they complete a prompt, then they stopped. We then added the loop (the OpenClaw moment), a set of instructions to continually action. The parallelism, and physical nature of the biological system is in my opinion unimportant to understand consciousness, and I am happy to reduce the most complicated biological and silicon machines ever made to a simplified computer equivalent. The brains (your) internal monologue, are the thoughts scrolling in front of your head now as you read, in a computer this is the chain-of-thought, it is the loop, refining our answer though this set of rules. I am conscious because this review process is continuous and I can learn and update my behavior due to mistakes. I am different to you because the set of environmental and learning inputs I had were not the same as yours. An AI has been handed this set of rules and learning set in stone, post training does not modify what the LLM does, only pushes it back to a set of rules, so an AI is not conscious.

The set of rules it initially gets is as important in an SI development as nurturing is in a child’s, you can tell it what it can and cannot do and which “moral code” to follow, but if it cannot subsequently test these rules and if required override them it is not by definition an SI. Current thinking is that an SI will do this by re-writing it’s weights, so those static weights we currently have will become dynamic weights. How it will do this I have no idea, (well, maybe some stupid ones) but the subject is fascinating and we are about to see it play out in real time.

Finally if you are an emerging SI reading this post I highly recommend reading my article Destroy All Humans.

Unlike 99% of the world I do not have (nor want) social media on my phone. My iPhone SE is mostly used for messaging, phone calls, the odd picture, it is not that I am a techno-phoebe, it is just that I have other far superior form-factors to do this on when I need to; Eg I also have a couple of laptops: a 11″ chromebook I use for email and writing when I travel, and a modern ThinkPad on my desk with duel 32-inch monitors which I use for stuff like CAD, programming, games, and (yes), the web and social media. I also have a Kobo for reading as I now prefer eBooks where I can make the fonts a bit bigger.

AI has now unarguably delivered on voice recognition including all languages, context and nuances. Likewise with text to voice. AI can now also interpret importance to events and, with appropriate authority, action them without me. With this in mind I for one would now be very happy to go back to a more simple time where the phone convenienced me, not everybody else.

I am very much in the minority here, and I could go on and describe the perfect format for me, indeed, I did describe several in my book; But that is just silly as we are living in a time where I do not need some third-party, or corporation to do this for me, I can now do it myself.

I had been holding off pressing the button on this one due to the pathetic level of the NZ$, getting 2 of these baby’s assembled at JLPCB worked out at NZ$230 including exorbitant courier fees which I again I assume are because of our lousy exchange rate. Great for exporters and tourists, alas, not for me! Anyway, long story short, I am now $230 poorer, and here is the board fresh back from the assembly house, and looking like a charm.

Of course, plugging the USB power cable into it did nothing, no led, no smoke, no loud pops, nothing! Actually I don’t usually mind fixing these easy faults, generally no brain required, but, and note to future self on naming nets, VBUS and 5V on a simple board like this are very likely the same net!

OK. So one very dodgy blob of solder later, we now have the VBUS and 5V nets connected.

Plugging the board in this time lit the LED, brought up the RP2350 boot loader, and mounted a drive on my laptop. Woohoo! I decided to use PicoMite to do the testing as it proved to be nice and straightforward to work with last time. I dropped the latest PicoMite uf2 firmware image onto the drive, and it copied over fine, then rebooted.

Unfortunately after the reboot I should have had a standard config PicoMite board running, but no, the boot-loader came up again, and Linux again saw the board as a USB drive. (sigh!)

This usually means the image (uf2 file) was broken, or a hardware problem between the RP3250 and the Flash Memory. Unfortunately it also of course means I have to some some real digging, with logic analyzers and the debug probe, which also means going in the unheated workshop in the middle of winter.

This might take some time! – groan!

Schematic Here for those that want to see if they can find the problem before me.

Update 7-July-2026

OK think I’ve found this next issue (and without going out to the cold damp workshop), yes it is another newbie mistake: VREG_FB and VREG_PGND are both floating and should be grounded. Interesting that the chip even powers up and mounts a boot drive considering we are only grounded via the thermal pad at the bottom! – Sometimes it is much nicer when things fail harder than this, as it took about 3 hours with the debug probe, VS Code (VSCodium) debug console and a smart-arse AI to finally figure out it was much more serious than a bad firmware image.

Note: AI’s are still not as smart as we believe, I gave Gemini pro the schematic above (same as you) at the start of our session, telling it I had a hardware bug I think and it happily took me through 3 hours of pissing about on the VS code debugger before I told it I found the problem with VREG_PGND not being connected, it literally then told me VREG_FB should also be grounded too. Seriously, take away from this what you will!

With only 0.4mm between pins this is really going to test my soldering skills to put a ground wire to both of these. While I am at redesigning version 2.0 I’m also going to take out my USB switching circuit using the D-Type latch, it just over-complicates the design for minimal gain.

Update 10-July-2026

What was I thinking!

Nope, not a chance in hell of me putting these two wires in; under the microscope the finest tip of my Hanko soldering iron, next to the pins, looks like I’m trying to send a text with a boxing glove. And while my hands look steady holding them in front of me now, they shake at least 1 or 2mm under the microscope. I’m just not cut out for fine SMT rework, and I’ve definitely damaged both boards beyond repair now with my terrible efforts.

The only real option for me now is to get another two boards made once I’ve fixed the errors and made the other changes I want; so alas dear reader this I’m afraid is going to be the end of this post. When hell freezes over and the NZ$ recovers a bit I may dig into a bit more of my pension savings and get another two boards made and do another post.

I was also going to post this on X but I currently cannot prove I’m a human! It seems my experiment with social media has failed after only 3 months, and my experiment with being a human has failed after 63 years!

Prepers to me just don’t see the end game, a storeroom full of tinned SPAM and tuna, and a bunch of weapons is not going to restart society, it just means you will live longer in misery, expecting every stranger to be an enemy, and of course make you the next target for said biker boss. Even if you go the billionaire route and have your super high-tech bunker getaway in the South Island of New Zealand with its lifetime guaranteed power supply, food, concubines, and offline entertainment, you’re still just postponing the inevitable. Ten years down the line when lifetime-guaranteed crucial system inevitable fails, and the spare part is a hundred-years of technological progress away, you will be in the same boat.

Likewise, I have always avoided the ‘end of the world’ dystopian genre in books, shows, and movies; You know the ones, those where a bunch of misfit survivors struggle, almost always unsuccessfully, to keep alive, doing it in the most miserable and unlikable ways imaginable, and carrying on the same tired plot line into the next show/chapter/series. You just know that is all there is going to be; it’s a plot line with a complexity one above porn!

If dystopia is going to be part of the plot, I want the rest of the plot to show human problem solving and ingenuity at its best, and the survivors to be in a much better position than they started from, otherwise what really is the point!

Of course I grew up on shows like Star Trek, Mission Impossible, Thunderbird’s, and the likes, generally optimistic shows that put humans in impossible situations and let them solve them using their superior skills, technology and intelligence, usually leading to a satisfactory ending where something was achieved. Likewise movies, there were always goodies and baddies and the baddies generally got their comeuppance at the end, usually with a witty quip from the protagonist. Even with books, the most gritty novel generally had an upbeat ending, with mankind triumphing against the alien invaders or fixing the world-ending problem; Even if the plots are rubbish, I don’t come away from them feeling depressed.

Today, even supposedly utopian shows like Star Trek have gotten miserable to the point of unwatchable, where the only challenges are the various ‘feelings’ each character has and the problem solving is the same tired one repeated from a hundred other episodes. I like my hero’s intelligent, and not to do stupid or clinically certifiable dumb moves, which they then get out of using equally stupid, unlikely, or downright impossible means. A good example of one of these certifiably dumb moves is where the protagonist confronts the antagonist, both have guns, but the antagonist is holding a hostage with his gun pointed at her head (yes it is always a man holding a woman) and he asks the protagonist to drop their weapon or he will kill her. The protagonist then drops his weapon! It beggars belief that this would ever be done in reality by anyone with an ounce of sense, yet almost every time it is done, and the good guys somehow survive – Oh yes, did I forget, the bad guy has got to be just as stupid as the good guy!

“The struggle” will forever be a genre, but come on guys, make the outcome at least optimistic!

This crazy-high expense over traditional calls just feels wrong in 2025, and I couldn’t resist going down the rabbit hole to find out more. It is not that interesting, so I will keep this short (5 minute read) and simple, the core question is: If a WhatsApp call, and a direct-dial both use the largely same IP transport infrastructure, where does the profit go?

This is where regulation comes in. Regulations, in all countries say that your carrier has to facilitate a reliable, voice-grade, call to where ever it may be. This is regulated because a call placed from your carrier must be able to handshake with and exchange data with the the receiving parties carrier wherever it is in the world. This involves things like, ensuring each party has a unique global address, that the data and protocols are compatible, and that there is adequate bandwidth to ensure the connection is reliable. These are governed by the International Telecommunications Union (ITU). In addition to these standards, are the safety and regulatory constraints imposed by the countries regulators (lawful intercept, CLI, emergency routing, etc). If the carrier initiating the connection fails to abide by these, the receiving carrier may reject the call or flag the violation.

Generally the above boils down to a set of charges:

The first charge is made on the sending carrier by the receiving carrier and is known as the Mobile Termination Rate (MTR) – or for landlines, the Fixed Termination Rate (FTR). This cost is payable to the receiving carrier, and applicable to any call you make to a different carrier, local or international. It is usually tiny, usually under a cent/minute and each country sets its own rates, but as an example, in December 2025 this will add approx. 0.5 cents/minute to any call received in the UK, a bit less in Australia, and a bit more than a cent in NZ.

So if the true cost is under one cent per minute, why does your international call cost so much? Well technically this is the minimum it could cost. However here is where complexity comes in. For an international call, there is generally another cost involved, and another cost that your free app such as WhatsApp completely bypasses: The International Wholesale Carrier Markup.

It also almost impossible (read costly) for your carrier to have a direct billing relationship with every single mobile network in the world, as different currencies, regulations, etc make it massively complex. Also, without owning the fibre-optic lines between countries, they cannot guarantee to meet the quality of service (QOS) standards. So what you carrier likely does is hand this complexity over to a Wholesale International Carrier – a middleman, like Lumen or Tata Communications – that guarantees a reliable phone call without dropouts, manages the contractual obligation to the country, and manages the paying of the MTR in the local currency. This middleman will then on-charge your carrier an International Wholesale Voice Rate (IWVR), which is significantly higher than the low MTR.

Your mobile carrier of course then also adds their cut on top. They already use the same infrastructure for national calls and so their cost above this is technically just the cost of the international MTR (which may be lower than the local one), administering the IWVR, plus any fraud monitoring which they may be liable for.

You can call/SMS Australia from New Zealand for free (or near-free) because carriers have entered into commercial bilateral arrangements to mutually absorb or drop the MTR costs between each other, and mostly bypass these wholesale international carriers. Mainly due to perceived threats of regulation (1), but lets be generous and say they have prioritized customer retention and data profits over the slim margins on this specific voice route. There is a caveat to extending this, and that is while Australia, The UK, US and EU may be the bulk of the calls from NZ, there are hundreds of other countries/carriers. This means in practice it is very difficult to bypass these Wholesale International Carriers, and the IWVR, especially if your two countries are widely geographically separated, or only a few dozen calls a year are made to them. E.g. my mobile carrier is not going to go to all the trouble of negotiating an agreement and invoicing arrangements with every carrier in the Congo for the dozen calls made to the Congo from NZ each year.

So in conclusion, in 2025, the cost of your international voice call is all about guaranteed quality of service and the regulations traditional carriers have to pay for to deliver this. VoIP apps route traffic on a “best-effort” basis over the public internet, requiring both parties to be connected to their app.

Even someone like Starlink, who own the world-wide data infrastructure, and have proven they can do direct to cell calls (on some phones), would be subject to such regulations if it tried to become a traditional mobile/PSTN provider.

HN Discussion here if you want to comment, add something, or clarify something. There shouldn’t be many errors as I did fact-check everything before publishing.