IndeanCondor

AI can write binary and probably better in a few years.

Assuming you mean an LLM, you'd have to train this LLM entirely on a parameter space of binary tokens. Or are you saying the LLM generating natural language tokens is going to be printing machine language in 3-5 years, because that claim kinda betrays a misunderstanding of LLM functions.

Probably the most head in the sand mil-tech article I've read in long time. There are very few explicit claims the author makes about Chinese capability that can be addressed, most of it just gesturing at anecdotes or vibes.

The author does not notice that the F-35 is a single engine jet rated for VSTOL flight characteristics. The F135 is required to produce that much thrust by itself to support that profile. The J-20 is a twin-engine fighter. Why, pray tell, does the engine designed for a twin engine, land based fighter designed for carrying large payloads of air to air munitions need to beat the thrust of the F135? The comparison is worse than stupid.

The Chinese Flanker fleet is being built out and maintained at scale with WS-10s, it's industries churn out 100-120 J-20s per year, all with twin WS-15s. This is a mature jet engine capability, at massive scale. "Not made in China"???

The author makes a passing comment that the WS-15 is "outdated" compared to NATO forces. They are clearly blissfully unaware that the F-18 Super Hornet standard runs the F414 powerplant, as old as the WS-15, itself an upgrade on the F404 powerplant, 50 years old now. The F-18F is the USN's mainline pacific theatre fighter.

I honestly believe anyone who considers the Chinese jet engine program to have been a failure to have perhaps lost some marbles along the way. It demonstrably is not, unless you think the PLAAF is about to collapse midair, a notion their daily ADIZ violations and interceptions over the SCS and the Taiwan straits should thoroughly disabuse. My prediction is by 2037 the entirety of Chinese domestic civil aviation will be running the C919 and they'll be a serious competitive threat to Airbus and Boeing.

Good lord.

a) If a single Portuguese treasure ship was laden with cargo of such magnitude, exactly how do you propose this is characterized as a TAM "discovered" by the East India companies?

b) Forget the asteroid, undersea mineral nodes are a total of 250 trillion tons worldwide, with an expected value of 233T USD. This is right here on Earth. Yet no country or company considers these minerals as a TAM, because the P/NAV makes this a net negative. This is the same reason any capex for Texas Shale oil will dry up if WTI falls below 60 for a period of time. No matter how many quintillions worth of oil are left in the shale reserves, if it costs more to extract than it will return, it isn't done. Arbitrary valuations mean nothing if they aren't economically viable.

c) Back to history. "Half the English crown treasury" sounds impressive, but the actual value was an estimated 500k pound sterling, upper bound 200M GBP in today's money. Quite a bit far away from 1 quadrillion. The English Treasury famously passed on a debt of 400k pound sterling to James 1 after the Spanish war, a contributing factor to a crisis between Crown and Parliament relations that eventually helped lead to the English Civil War. Suffice to say, a wartime treasury isn't very large.

Humour me for a bit. Let's assume we need to Hohmann transfer 1 quadrillion worth of minerals (30M metric tons) from a location in the asteroid belt to Earth. The DV necessary is 5000m/s, leading to a total energy calculation of 7.5 x 1e16 J. This is 78M tons of hydrolox fuel. 780k launches assuming 100 tons per launch (Starship). Are you honestly telling me that SpaceX is justified in adding asteroid mining returns to its TAM (1 quadrillion) because you believe it's economically viable to make 780,000 (lower bound) separate launches of fuel payloads from Earth to this asteroid? If ISRU is your claimed solution, then where is the Madre De Deus of this ISRU demonstration?

This is such terrible history. Please Google things before making analogies.

The East India Companies (Dutch and English) didn't invent or discover the spice trade. They were created to leverage private capital to wrest control of the trade from the Iberian Union, an entity both Company's sovereigns were at war with.

The "TAM of the emerging markets" had already emerged both because of the Silk Route from Pax Mongolica and in the 1500s from the Portuguese Empire's Cartaz trade license monopolies enforced by their naval posts (feitorias) throughout Asia.

The TAM was incredibly apparent, Anglo-Dutch privateering during the war had seized multiple cargo laden ships along with trade route information.

This statement needs qualifiers.

Are you claiming you have a raw binary to Fable and it just reverse engineered it by reading it? Or are you claiming (like for every other model released in the past 1.5 years) it's using an integration with Ghidra or BinaryNinja to assist - in which case I completely disagree even a 30B model can do that with those tools.

Also an FYI, AI advancement and Anthropic are not synonymous. Someone asking Anthropic to back up their claims is not coping about AI, especially as independent benchmarking of Fable is giving equivalent or slightly above par results to GPT 5.5.

The system card does not use any of the benchmarks used in the previous Opus 4.5+ system cards. All the scores are in Anthropic owned benchmarks. I find it extremely hard to believe the marketing department of the company was not involved in a material release to the public - which is the marketing departments literal job.

Based on all I know about Deemed Exports wrt software and current US controls on software Deemed Exports, your read is spot on.

The phrasing of the foreign nationals implies a Deemed Export control, which is already in place for software for stuff like drones or space satcomms.

If it's a Deemed Export control, it's a strategic position and not a knee jerk reaction about cybersecurity threats.

It's a coherent read too; if Fable can solve coding and build biological weapons (X to doubt) - well then terminal guidance and autonomous drone controls should be a piece of cake for it and that software is already under Deemed Export restrictions.

It was a coalition of monarchies, so terrified of the prospect of even a single popular republic rejecting the divine right to rule, that began the bloodshed and invaded France.

Blaming the Revolution for the Coalition wars is just bad history.

In case you didn't notice, the Revolution won. All contemporary republics fundamentally inherit from the French Revolution - you're surprised that the systems of govt honour and romanticize it's progenitors?

Aside, if bloodshed prevention is your only barometer for history, supporting the Coalition view of events is even sillier, because the Bourbon Restoration directly led to the Revolutions of 1830 and 1848, which was yet more bloodshed. Almost as if unjust systems are fundamentally untenable...

And yet here you are, strongly reacting to a philosophical argument made by him.

He's mortal, he's dead, and yet he's affected you here, a throwaway comment on a forum post no one will think about in a week's time.

The same UK security research body ran the same CTF against GPT5.5. GPT5.5 got the same result as Mythos.

Anthropic promised us that Mythos was such an existential threat that it would compromise "every OS and browser on devices across the planet". They've held conferences and meetings with banks and govts across the world, shouting how critical this issue is.

GPT5.5 has been out for a month. Every device on earth has not been breached yet. It's very fair to criticize Anthropic's maximalist posturing when it's becoming exceedingly clear their models are fairly behind OpenAI's in capability.

In my opinion, the original commenter's statement stands, and the UK govt data point only helps support that due to the equal result between Mythos and GPT.

I'd advise reading into the specifics of what happened with Firefox; the TL;DR is a reduced safety version of its code was scanned by Opus 4.6 (yes Opus) and found a multitude of bugs and 4 high severity vulns that did not escape sandbox. The Mythos system card test describes running Mythos against the same issues Opus found to see if it could reliably replicate and chain together an attack.

Autonomous loitering munitions with 'AI' (image classification CNNs) are already in service and have been used - most demonstrably by the IDF.

Even during the Nagorno-Karabakh war, Azeri loitering munitions were able to suppress Armenian air defenses by hitting them when they rolled out of of concealment. I believe that killchain requires a level of autonomous functionality.