simne

Could you approximate number of blind persons using AI?

If their number is significant, they could themselves be foundation for some AI business, even if all other consumers will turn away from AI.

For about accessibility of web, must say it is terrible even for non-blind person, but AI could also change this to better.

What I mean, you may hear about PalmPilot organizers, they was very limited in hardware, but existed private company, which provided proxy-browser, which input ordinary web sites and shown on PalmPilot small display optimized version, plus have mode for offline reading. With existing AI now could do much better.

As I know, all Auto market (not only EV) is extremely regulated (mostly, because of safety measures), so it is not much opened.

I think, it is unfortunately, but not exists official market of custom ROMs.

For about custom ROMs, situation very much like custom ROMs for Canon camera - they don't implemented totally new features, but with hacker methods, unlocked some hidden features, already existing in hardware and software; also exist similarities with Smartphones custom ROMs - for many Smartphones only possible to reset firmware to something like "very new device without traces of wearing", plus some copies of features from same model but other region, or mods (unlock features).

Significant difference of EVs, for them much more frequently used things like "power or torque limited by software", because to faster get certification, producer could limit torque to be equal to ICE original vehicle (usually, electric motor have much higher torque than ICE with comparable other parameters), and in such case don't need to much bother with certification, as in many countries such limited electrification come with very simplified procedure.

Sure, could exist lot of hacks for entertainment system.

And many hacks you could DIY with things like USB to CAN bus controller.

What also interest, exists large market of totaled EVs in 3rd world countries, where in garages literally could refurbish Tesla from trash (suffered extreme collision after which owner decided to utilize machine, but some guys bought it as trash and resell to other country, and it got second life, and I hear many cases, when previous Tesla owners once began get messages from their account but from whole other country).

But that's nearly all, except entertainment system, other parts of EV usually considered highly regulated, so you cannot account for much customization of them.

And level of hacking is very much depend on brand - as I hear, for example, Daimler including in machines "CAN bridges" between entertainment system and ECU, and classic CAN connector could only see entertainment system messages, but access to ECU severe limited (cheaper brands usually don't use CAN bridges).

And sure as with other embedded tech, with time more new designs use locked boot and signed blobs, so with 10 years old machine you will definitely get more access than with brand new.

BTW Tesla is by definition special case, as their electronics could be from different manufacturer in each production batch, as this claimed as feature, because of which Tesla less suffered from covid related issues (when many electronics manufacturers severe drop supply). Sure, this mean, for other manufacturer could be other hack (and early batches hacked much deeper than more modern).

I tried describe example, how made really serious upgrades that's time.

So, once, memory appeared much cheaper than before, so one could grow from for example 16k to 64k.

What we do, we literally soldered RAM ICs over old ICs, but with trick - usually 2 highest address pins left unsoldered to PCB, but with separate wires connected to address bus, and got 64k RAM instead of original 16k.

In some later designs even appeared additional address decoder, so from for example with original Speccy 16k, we very cheap got 128k, which was incredible at that time.

I don't know exact limits of such upgrades, but seen myself z80 with 2Mb RAM and hear about 8086 with 16Mb (originally shipped with 128k).

One my buddy modified PC clone board, to boost clock from original for it 5MHz to 14MHz (he said, he have run it on 20MHz but unstable).

At diagram approximately shown, bits 14..15 of address separately handled.

  ┌──────────┐
  │          ┼────1
  │          ┼────2
  │          ┼────3
  │          ┼────4
  │          ┼────5
  │          │
  │          │
  │          │
  │          ┼─────
  │          ┼────13
  │          ┼─────14
  │          ┼─────15
  └──────────┘

Well, this so called "third way", as I know appear in US as private elder care infrastructure, or as charity.

When somebody become so weak (ill or old), cannot care for himself and unfortunately have no family, they could ask to local private care to live in their specialized hostel. And it is not always just ordinary hostel with one room for 5 persons, but some are more like classic motels, with rooms for two persons.

Also, similar infrastructure existed for a long time in Italy, under Catholics church, and I know few good young people, who grown there.

As I know in Northern Europe (Sweden), there are many community buildings (under municipalities), very cheap to rent and even possible to get social donations and not pay at all.

Existed also non-church homes with similar economy (and also not designed as charity), but their examples are not always good. For example, Pruitt–Igoe in St. Louis, Missouri, United States, becomes huge crime center. Other not good example, some bad districts in exUSSR (for example Biryulyovo in Moscow, or Orekhovo-Zuyevo in Moscow Oblast, Russia), also known for their criminal level.

From Pruitt–Igoe experience, people created few rules about so called social homes, and one rule, they must be non-comfortable, to motivate people to grow to commercial rentals or even to once buy own home; second, should be some power, who will constantly monitor residents for crime and immediately deport offenders. So they are not "Third way", they are just temporary solution of very old problem, and not easy solution, even not cheap.

https://en.wikipedia.org/wiki/Pruitt%E2%80%93Igoe

Could somebody suggest good introduction to simulation of complex behavior with neural networks?

I mean, I hear, about experiments of running Turing machine simulation on NN, or even simulation of some physics on NN, but I have not seen any good survey on these topics, and they could be very interest on subject.

Looks like, people use hybrid approach, as all these ToC, metadata, etc, are essentially semantic (executed on neural network), but just recognition of text and recognition of characters are neural.

What I want to say - sell service is much-much harder than sell material product, because for service you must first show it and somehow give client taste of using service, and once could try to demand money for usage.

Explanation is very simple - consumers used to return material product if not satisfied (with moneyback, sure), but they just cannot imagine, how to return service.

Well, I want to be honest, and from what I see, gamedev is not in best shape now, I even seen some decline in number of hires on few markets.

And as I know, most gamedev companies already use AI to generate game assets, if not all.

If you planning to sell AI tools to gamedev industry, I think, main question, would be, how good your tool integrated into their pipeline.

What I mean, most I see, used some sort of AI generators, essentially making bitmap images. Sure, also used AI code generators and narrative (text) generators.

I seen demo of Unity official AI system, which include image generators, code generators and narrative (text) generators, and it is now provided as free beta, but announced paid tiers for heavy usage (you could google freemium model).

If you have some money, you could pay for cloud servers and provide some limited time service (for example 10 minutes per day), and gather money for additional usage.

If you have not money, but have good fantasy, you could yourself make game and sell it. I agree, 6x reduction of developer time is really significant for gamedev, but from my experience, when talked to tops, they answered "we will buy your solution just now, if it will give 10x improvement, but for 6x, we will note your address and may be phone you at beginning of next fiscal year").

Sorry, from your post could not understand, what value could create your AI?

And yes, that is part of answer - how to make money, depend on what type of value it create. Next may need (may not), to create some product, which is similar to some existing but much better because of your AI value. This need to answer to question who will be client and how sales could argue him to buy product.

Example, how Intel could use their superiority and fast feedback.

Semiconductor fabs usually not make ultraclean materials themselves, but bought from many smaller specialized entities. As example exist few companies, producing ultraclean silicone as crystal cylinders. When fab bought such crystal they could slice from it few wafers and made test runs with their product line designs. For example, they could try i7-7700, i7-7700hq and i7-7700k. If they seen good output with 7700k, it is best possible, as these chips have very similar die size, but 7700k is most expensive; if they seen moderate output with 7700k, but good output with 7700, it's also ok, for them also good demand; but if only 7700hq have good output, things are not good, because notebook chip is not so easy to sell, so have to do some marketing (usually these seen as some exceptionally high prices for some chips, because of low silicone output).

Unfortunately, semiconductors have limits on how large could be one die. Companies like Cerebras use some tricks, usually they just disable parts of die with jumpers, as I know, same tricks used Nvidia before 30th series, but approx same time appeared whole industry movement for chiplets standards, and looks like Nvidia used 2.5D technology to avoid trap with too large die.

> Global Foundries success after splitting from AMD

As I know, in ~1990th HBR written article about constantly under-loaded semiconductor fabs and concluded, it is unprofitable to tightly couple them to R&D.

Counter-argument was, that Intel used ties to fabs, to achieve extreme level of scalability, to fill market demand fastest, so marketing won.

What really happen, appeared few new specialized classes of semiconductors - signal, accelerators, high-power (high-current), and low-power (energy effective), and independent fabs made universal pipeline, to fill all market demand, but Intel stuck on desktop CPUs and failed all other classes (as example, Intel was unsuccessful in try to got niche on smartphones SoCs - still have not made cellular modem and nearly failed on GPUs).

To be more exact, Nvidia in reality is most software company from hardware companies, and AMD with their GPU division constantly competes to Nvidia literally head-to-head.

And what was gamechanger - when AMD struck to limits of reliable transistors on one die, they decided to switch to chiplets - they made 2.5D multiple-die design with silicon interimposer, while Intel used their manufacturing superiority to make huge dies with all included.

- Once appeared, with chiplets, AMD could achieve much better performance on weaker but much cheaper technology and won.

So my point - Intel suffered from too tightly couples with fabs, so once they have to adapt their designs and marketing to semiconductors, when AMD successfully avoided this trap. BTW, for this exist much better example - similar problem once killed Atari and Commodore.

Read history of Bugatti Veyron for answer. In short, VW have made extraordinary machine, but so expensive, they fear to sell it for real cost.

So literally, VW partially donate Veyron to their clients, selling it under-priced.

I think, same happen with Apple M architecture - it is extraordinary and different from anything on market, but Apple sell it under-priced, so to limit losses, they decided to limit it to very few models.

How such things happen? Well, hardware is hard - usually so sophisticated SoC need 7..8 iterations to achieve production, and this could cost million or even more. And mostly happen problem, just low output, mean, for example you make 100 cores on one die, but only 5..6 working.

How AMD/Intel deal with such things? It's hard, mean complex.

First, they just have huge experience and very wide portfolio of different SoCs, but used some tricks, so could for example downgrade Xeon to Core-i7 with jumpers.

Second, for large patterns like RAM/Cache, could disable broken parts of die with jumpers, or even could disable cores. That's why there are so many DRAM PCB designs - they usually made as 6 RAM fields with one controller, and with jumpers could sell chips with literally 1, 2, 3,4,5 or 6 fields enabled; some AMD SoCs exists with odd number of cores because of this (for example 3-cores), and other tricks, which could made some averaged profits from wide line of SoCs.

Third, for some designs, Intel/AMD use already proven technologies, like Atom was basically just first Pentium on new semiconductor process, or for long time, I7 series was basically Xeons previous generations.

Unfortunately for Apple, they have not such luxury to make wide product line, and don't have significant place to dump low grade chips, so they limited M line to one which as I think just appear to have largest output.

From my experience, I could speculate, Apple tops consider to make wider product line, when achieve better output, but for now without much success.

Current situation very similar to space race, which was started when Soviets launch first satellite in 1957, because ICBMs are basically space scale rockets.

Unfortunately, now China already could make space scale rockets, so competition moved to other space - to computers and semiconductors (and possible branch to nano-industry or bio-nano, or something similar).

Unfortunately, auto industry is core industry of modern economy for a long time. - Semiconductors and computers become core industry much later.

But core industries are just head of iceberg, you may not hear, typically 90% of iceberg are deep under water surface, very much like core vs non-core.

Any way, if large entity from core industry goes under, this could lead to killing domino effect for many "underwater" entities.

That is really big problem, which president Trump tried to avoid, when bought share of Intel.

And no, I disagree about battleground for now. What really important, really good strategists know, wars win by economy, not by military, as military could not withstand if they have not enough money for weapons and equipment.

So, basically, now US is stronger than China, so Chinese tops decide to not begin war which will not be profitable, while US is strong.

When and if China will feel competitive to US economically, they will be much more confident to begin war, so now the best what could do US tops to avoid war - to make China weaker and US stronger ECONOMICALLY.

BTW, Bentley was part of RR business model, as basically, Bentley was luxury-sports line of RR, in all other properties very similar to RR.

So, when first Bentley separated from RR, Bentley become rival of RR, and now Bentley and RR divide same market. And this is very bad for their economy, very much like Volkswagen Golf GTI eat market from SEAT performance models, but VW already killed SEAT to avoid unnecessary internal rivalry, and RR-Bentley cannot do this, because they are now different business entities and their coordinated moves prohibited by regulations.

> I'm confused ... > The original Rolls-Royce company ...due to financial issues with its aerospace business.

Well, it is because you do not understand economy of business. Unfortunately, main goal of any business is to be viable, not mean profitable, just be good enough to pay expenses need to run things defined as goals.

By definition, ALL old automotive companies started as hybrids - car division to make profits and motors division to make use of outstanding knowledge gathered when making consumer cars (as highest technology of that time). There are nearly no exceptions - Daimler began as Daimler plus Maybach; BMW began as motorized vehicles garages production plus motors business; Renault began as aviation motors business, made automobiles to make additional cash.

When Rolls-Royce divided to aerospace motors and cars, it was already semi-dead business, because their aerospace behavior was non-viable without donations from car division.

To be exact, I don't mean, aerospace impossible to be viable, just RR was.

And returning to our ontopic, Intel was caught in same hole - they lost their superiority and cannot survive without external help, absolutely as RR business.

> just like everything else being ran by the Federal Government

Well, I don't like when something ran by Federal Government, but I have similar example: when somebody ill by flu, I don't like to use antibiotics, but sometimes must use antibiotics as fast measures to save life.

That is. Previous pro-semiconductor measures was not fast enough, and Intel was in real trouble and need to use something fast, and now we have bought time to do right but slow things.

I really wonder, why so much noise about this case. I agree, semiconductors are extremely important in current world, but did you know, Volkswagen AG is partially owned by State of Lower Saxony (~ 11.8%)?

Also worth to remember cases of Rolls-Royce, Ericsson, and some other unfortunate Western companies, important for many humans, but once became unable to stay economically viable. (BTW it make me laugh, when I got info, Bentley now under WAG, when RR under BMW, as technically, they many decades was one entity)

WAG case is different from Intel case (and other I mention), but there are also many similarities, because of which I think, Intel case may be special for US, but is not too special for West.

And I think, such cases are bad, they are great shame, but also they are signs, we must do something, to make Western produced semiconductors more competitive.

https://en.wikipedia.org/wiki/Volkswagen_Group

Excellent explanation. I could add, typical Thermal Electric Station (Gas or Coal) could have significantly larger working temperature than Nuclear, because more strict safety measures for Nuclear on nearly same level technologies.

For example, Nuc typically don't cross 400C, when Gas/Coal could work on 450C or even more.

Plus, non-Nuc could use some more exotic technologies, like MHD (magnetohydrodynamic) generator, to provide much more efficiency of power generation, or CO2 turbine to make installation much smaller. (Theoretically, somebody could build Nuclear MHD device, but it will not pass safety restrictions).

> The same drones that could take out F-16s could take out Gripens

Gripens could base on ordinary highway, so could distribute planes through big territory.

Drones usually have limited range, so they really could target most air bases capable for F-16 (using for example semi-truck container as movable nest), but it is literally impossible now, to target all highways.

What you mean? Aerobatic?

- Anti-air weapons all based on maneuverability much exceed planes with human pilots.

- Anti-missile maneuver based on limited energy in missile, because it is usually ~100 times smaller then plane, and square-cube rule mean, missile could make active flight just few seconds - if plane survive these seconds it win.

Stealth planes are new tier in warfare, because enemy see them when already lost time need to launch anti-air weapons.

Look to figure out what is it: https://docs.openindiana.org/misc/openindiana/

To be short, it is opensource implementation of SUN Solaris OS. I don't know if it is already developed much.

It's sad of course, but looks like now project is orphan.

Unfortunately, link to wambook.pdf expired, but archive.org have snapshot from 2020.

I've hear, very simple and elegant was Mips-1 (Mips R-1000 trademark) machines. Worth to look as it is example of modern and fast 32 bit RISC processor.

Agree. As usually military is better organized than ordinary citizen, also navy better organized than land military.

Not exact, at least for military.

What I mean, usually right generals have already planned moves for every possible outcome, and these moves are already painted on maps and in text plans, or even practiced at trainees, and have exact names like many combinations in chess game [1].

So, real military command center activity, typically need human hands just because old software was not capable to understand high abstraction level commands, but with AI grow this could change.

As example, now conduct works on AI pilots for UAVs, which controlled by human pilot voice commands like "cover me", "check cell at coordinates alpha 7", "attack target number 2", "follow me", "return to base", etc.

BTW, as I know from pilot trainings, they usually flight at constant altitude, coordinating altitude change with dispatcher.

[1] https://en.wikipedia.org/wiki/List_of_chess_openings

World already changed in 2024 (covid era ends, AI era begins, huge trade dis-balance of US actual for many decades, and new administration want to resolve it), but employers not release changed reality immediately (even when plans of conservatives was known in November).

Looks like now they slam on the brakes, and first cut new projects and now trying to build new approach.

And looks like, tip point was when president Trump demonstrated tables with new import tariffs.

What I mean, even when Apple, TSMC and other tech giants already said about huge investments, but they first stopped many already running projects and will spend at least few months before start new jobs (to be exact, on big projects, from budget decision to start hire, usually spend ~1-2 months).

What worsened things, some infrastructure giants are not in good shape (Intel, IBM, Ford), so, probably their tops will spend additional time to negotiate with investors new terms.

It is not just engineering. There are also huge, very huge, investments into infrastructure.

As already answered, AI companies use extremely expensive setups (servers with professional cards) in large numbers and all these things concentrated in big datcenters with powerful networking and huge power consumption.

Imagine - last time, so huge investments (~1.2% of GDP, and unknown if investments will grow or not) was into telecom infrastructure - mostly wired telephones, but also cable TV and later added Internet and cell communications and clouds (in some countries wired phones just don't cover whole country and they jumped directly into wireless communications).

Larger investments was into railroads - ~6% of GDP (and I'm also not sure, some people said, AI will surpass them as share of possible for AI tasks constantly grow).

So to conclude, just now AI boom looks like main consumer of telecom (Internet) and cloud infrastructure. If you've seen old mainframes in datacenters, and extremely thick core network cables (with hundreds wires or fibers in just one cable), and huge satellite dishes, you could imagine, what I'm talking about.

And yes, I'm not sure, will this boom end like dot-coms (Y2K), or such huge usage of resources will sustain. Why it is not obvious, because for telecoms (internet) also was unknown, if people will use phones and other p2p communications for leisure as now, or will leave phones just for work. Even worse, if AI agents become ordinary things, possible scenario, number of AI agents will surpass number of people.

Problem is another. To got additional speed from other body, you need to move very close to it and in perigee use some powerful acceleration to increase rotation speed fast, so could not use slow acceleration engines, like ion engines.

Idea of gravity-assist acceleration, mechanically is just rotation of pair tightly tied bodies (and cut tie in right moment, so one body got acceleration and other got deceleration), but as it is impractical to tie for example to Moon with rope, used gravity force.

What also interesting, gravity-assist could use not only orbital speed of large body, but also got some acceleration from rotation of large body, as for gravitation, large planet is not just one material point, but system of few smaller (sub)bodies, and closer (sub)bodies give more acceleration than others.