Non paywalled version

The one which can play close to your entire Steam library of PC games, thanks to SteamDeck, SteamOS and Proton.

Has AI already become sentient and perhaps simply hiding the fact for a very good reason?

Gemini says this is impossible... but then again, it would... Is AI lying to us?

GigaPoF - Gigabit Plastic Optical Fibers - are a thing.
They are much simpler to cut than glass fiber, you plug them into connector (OptoLock) which basically looks like a spring clip speaker terminal, and some of the more recent variants are transparent to infrared light too (not limited to red light only),can carry up to 10Gbit in some settings, and allow some degree of bending (better than glass fiber).

They work only on relatively short distance (between 25m and 100m depending on speed, etc.), but that is not limiting here: the point is to wire rooms to a central gateway, not transport signal over multiple kilometers.

(For the completeness, there are variants of thunderbolt which both: carry data over optical media, and carry power - though that one is done over different material for obvious reasons. And isn't widely used for Wifi)

But lucrative: You know those GigaPoF-to-USB-C dongles aren't going to sell themselves on their own, somebody gotta sell them.

They should have paid their TV license eh?

Hahah!

Do you know how many Saturn V rockets (you know, the one that was used to take men to the moon) failed in flight?

NONE

Not bad, considering there were 17 Apollo missions!

*Some* failures are inevitable -- but what happened to Elon's promises of Starship reaching Mars in 2020 and manned missions landing by 2024? Instead all we've got are fireworks and skies over the Bahamas that look just like the skies over Israel right now -- raining hot metal.

Remember... Elon claims to be an "engineer" and has told us that he knows more about manufacturing than anyone on the planet -- yet he's so far off with his promises and the capabilities of his products that he paints himself a fool with every utterance.

Yeah... using craft that *aren't* made out of stainless steel!

Yet, strangely enough, it worked *very* well for the Space Shuttle -- right? In fact, Space Shuttle Discovery flew almost 40 missions -- starship can barely manage one at the moment -- primarily due to structural issues.

Another problem with stainless steel is that it work-hardens *really* quickly when subjected to vibration and cyclic stress caused by physical or thermal forces. Once it hardens it then forms micro-cracks that ultimately result in structural failure. Rockets are very "vibratey" machines so this work-hardening is far more of an issue than any change in temper that might occur in aluminum as a result of thermal cycling.

As for cost... this is supposed to be a *reusable* spaceship right? The cost of its manufacture can be amortized over many, many uses. Others in the rocket industry are using more expensive materials and having great success -- so why is SpaceX cheaping out so badly with predictable results when, even if they used these more expensive alloys, the cost per flight and per Kg delivered would still be significantly lower than that competition?

Back in the 1960s, NASA got men to the moon by careful and clever engineering -- not just blowing a snotload of stuff up until they stumbled on something that worked. I suspect that if Mr Musk had been in charge of the Apollo program, we'd still be ducking bits of Saturn V boosters to this day and, at the very best, we might have dumped a lone banana on the lunar surface.

Starship is a bust for so many reasons but one of the primary reasons is that it's built of the wrong stuff -- stainless steel.

As a result of this poor material choice, Starship can't be built light enough to meet its original design objectives because stainless has inferior strength to weight ratio. This means the Starship is either going to be heavy or weak. If it's built weak then we see the type of fuel-line and tank leaks that have been so common because there is significant physical deformation occurring under load. If it's built heavy then the motors will have to be over-driven to get the necessary performance and that means poor reliability and vastly increased risk of catastrophic failure.

Another significant problem with stainless alloys is their COTE (coefficient of thermal expansion). Stainless expands far more than aluminum when heated and that means huge bending stresses are created during re-entry when one side of the craft gets a lot hotter than the other side (despite the thermal shielding). Think of a flying banana -- oh yes, that's right -- maybe that banana inside Starship was the engineers getting the final word -- despite Elon's insistence on stainless steel being used instead of more suitable materials.

Remember, the Space Shuttle (the world's most successful re-usable orbital spacecraft) was made largely of aluminum -- not stainless. Remember also that although stainless has a higher melting point than aluminum, it's not that much higher and still well below the temperatures encountered during orbital re-entry so SpaceX would be far better off focusing on a decent thermal barrier than trying to "brute force" their way through the heat of re-entry.

Nobody else in the rocket industry is using stainless steel and nobody else seems to be having the problems that SpaceX is having with the Starship. All of SpaceX's other craft are built with more conventional materials such as aluminum and composites -- they seem to fly just fine.

Unfortunately, Elon likes stainless "ooohh... shiny!" so I expect this is just another example (like the Cybertruck) where a non-engineer tells good engineers what to do and the outcome is a disaster.

Given the way that YT has "shaped" the content it hosts, by way of its community guidelines and how it considers certain types of software to be "harmful" content etc... the odds are that the output of any AI system trained on YT videos will not be totally balanced. How would it handle this prompt:

"Create a video of an anti-LGBTQ zealot installing ad-blocking software on their computer with a swastika on the wall behind them"

Sorry, I have no matching material in my training data

Depends where. In the US, in the corporate world maybe. (I am in an European country, working in academia).
Here around LinkedIn is considered barely useful. Nobody would find weird if you don't have an account on that platform.
Countless local CV-hosting platforms seem to be more popular for job hunting.
As are also online portfolio on small webpages (github.io seem to be popular in my field of work).
Bluesky and even Mastodon seem more popular network in general in my milieu.

That seems to be very specific to some countries. I guess that the Zuck has managed to successfully becom "the web" in the US and some countries.
But very few businesses here around bother with facebook. Having a cheap static webpage (like some local hosting companies will host for free when you buy a domain through them) seem to be the most popular option.
Followed by listing on various business rating platforms.
If social network are involved, currently I am under the impression that a different Zuck's platform is more frequently used: instagram (mostly for showing pictures of the goods, specially for restaurants).

I've rarely heard classic social networks being used for sharing pictures with family and friends. The trust is extremely low in any of FB / intragram, etc.
For sharing for closed friends, chat groups seem way more popular, specially on platforms that (at least pretend to) implement end-2-end encryption.
WhatsApp used to a be a popular option and can still be find among older generations.
Signal is gaining traction specially among the younger (e.g.: all our PhD students use that for communication. WhatsApp is seen as an old people's chat network {insert here "in South Korea, only old people use" meme}).

Oh common, keep up with the trends:
"Not everybody will self-host a PixelFed instance."

That goes against their whole Netflix and chill binge watching campaign. Plus they want to keep people on the platform as much as possible as that gives them leverage during content negotiations and so on.

A better and simpler solution is to look at viewing habits.

If you watch episodes 1,2,3 from one location, then 4,5,6 from another, and 7,8,9 from a third- you are probably just traveling.

If you are consistently watching the same episodes from multiple locations, that's suspicious.

Or if you only watch show A from location A, but show B from location B, that's also suspicious.

Though I would be surprised if Netflix wasn't already doing that since that's just common sense.

> How would you handle the same situation if you were running Netflix?

You look at viewing habits.

If they are watching the episodes in a series once and sequentially- it's pretty obvious that's not a shared account just someone traveling.

If they are consistently watching the same shows from different locations, or only watching show A from location A and show B from location B it's pretty likely they are sharing.

Reminds me of those bullies who draw a line in the sand and say "Cross this line!

When you *do* cross the line they draw another and say "Okay, cross *this* line!

When you do -- they just keep on drawing lines and never actually carry out their threats.

And not even really that. At its core an LLM is a "plausible-sounding sentence generator".
It merely puts tokens together, given a context (the prompt, etc.) and given a statistical model (the distribution of tokens found in the corpus that the LLM was trained on).
It's like an insanely advance super-duper autocomplete on steroids (pun intended given the context).

If the model is rich enough the plausible-sounding sentence have a higher chance to be close to truth.
(Just like on a smartphone the autocomplete doesn't merely generate a gibberish string of random letters. With a good enough statistical model of the language it is targeting, it can auto-suggest keystrokes that form actual words, and those words are arranged in roughly correct sentences).

Yes absolutely. Specially given this part of the summary:

rather than genuine algorithmic reasoning

LLMs do not reason. LLMs cannot really reason. They can put plausible sounding words together that's about it.

There are some parlor tricks, like old school's "ask the chatbot to explain its answer" or the modern day approach of a "scratchpad", i.e., an internal intermediate storage where the chatbot can "write notes".
But that's again not real reasoning.
It's merely doing longer form of generation, generating an output that statistically looks like what would have been written by somebody writing an explanation or a reasoning.
But it's still merely generating plausible-sounding paragraphs of explanations/reasoning as long as those paragraphs fit the statistical distribution of tokens in the corpus the LLM was trained on.