Dammit! In trying to link to it, I just discovered the wonderful site wisdomofchopra.com is no more. Probably because Chopra recently expanded his grift from books and seminars to an online AI, so the old 2012 random sentence generator site was competition. Oh well. Here's the WisdomofChopra Twitter feed: https://x.com/WisdomOfChopra
Which reminds me, inspired by Deepak Chopra's sayings, actual philosopher Daniel Dennett, who was known for saying things which were not only coherent but often quite insightful, coined the term Deepity: "A statement that is apparently profound, but is actually trivial on one level and meaningless on another."
This is a good step. It's unfortunate it came after significant user blow back from rapidly deploying AI features without surfacing these controls.
I hope the Firefox team learns from this because a vibrant, healthy Firefox is vital as an alternative to the browser engine duopoly controlled by commercial tech giants.
Well said. When I read stuff like this implying 'consciousness studies' can be a hard science, the undefined and ill-defined terms just keep piling up until I nope out. I'm lucky if I make it past the abstract to maybe the third paragraph. No reproducible science can be built on so many subjective vagaries.
I think consciousness studies can be interesting but they need to stay in the philosophy department between Searle's Chinese room and the P-Zombie lounge until they're ready to experimentally test falsifiable hypotheses with neuroscientists. And until they have a rigorous, consistent definition of what human consciousness is and is not, they really need to stop pretending AI has any relevance to human consciousness. There's no evidence AIs are conscious, and even if there someday is reproducible evidence - there's no reason to think it might be similar enough to human consciousness to make useful predictions about either (and that's assuming humans are conscious, which is still a matter of some debate in the field).
Another poster found grabs of the company's website on archive.org. The last date it existed to be grabbed was in 2001. I think the OP is fine on this one.
Other than that, there's virtually no mention of the company or software anywhere online. Just to put that in context, I'm in the vintage computer / software community where thousands of amateur historians and archivists scour obscure corners looking for old, unknown software apps to preserve. Software sold for Windows 98 up to 2001 (so recent as to barely be considered 'vintage') with so little online footprint means it must have been incredibly obscure. No ads or reviews from magazines or even newsletters means there's a good chance it was a one or two person part-time, home-based business and the product had hundreds or maybe even just dozens of users.
1998-2001 was the hottest time ever for PC software. I worked in marketing Windows software during this period. To have any commercial Windows software product actively available for sale in the late 90s with no surviving footprint would almost require intentional effort to stay unknown. No press releases mailed. No review copies sent. No shows or conferences attended (exhibitor listings are searchable online now). There were much older niche vertical software programs for much more obscure platforms which we know sold less than a hundred copies ever, yet still have a larger online footprint than this program. The OP de-protecting and archiving this previously unknown commercial program represents quite a notable find in the preservation community.
They could have been bought by a bigger company. You never know until you get sued. If the copyright isn't registered, and often even if it is, there's no way to know who owns it now.
If you're legally daring, and you get sued, you can try to force them to prove they own the copyright. There's often not enough documentation of this sort of thing. People have gotten away with not paying their car loans because after several loan sales and company mergers, the company that owns it can't prove they do, so effectively nobody owns it and it doesn't exist.
Sure. Anything's possible. My point was that any company response in this particular case is vanishingly unlikely.
> They could have been bought by a bigger company.
Any acquisition (or bankruptcy) in the 2000s is highly likely to have created a web trail.
> If the copyright isn't registered, and often even if it is, there's no way to know who owns it now.
But trademarks need to be renewed, which creates a web trail. And companies that sue over things, have generally already filed suit (or been sued) at least once in their 30+ year history (assuming they still exist today), all of which creates a web trail.
Over my decades in the tech business, I've been fortunate to work with some excellent lawyers as my in-house general counsel and one of the best taught me an important meta-point about realistically evaluating the likelihood of getting sued. Because, no matter how unlikely, anyone can sue over pretty much anything. So after doing the likelihood calculation, remember to ask: "What would the plaintiff's damages be?" Let's imagine someone is crazy enough to spend many months or years and tens of thousands of dollars suing over this defunct, unknown software which hasn't sold a copy in 25 years. But, the law being the law, they win a judgement of damages amounting to a symbolic $1. Because the plaintiff has to prove the actions of the defendant caused them to lose some significant amount money - in the present tense. And making significant amounts of money selling software in the 2020s leaves a web trail.
Because crazy people exist, this bonkers $1 scenario has actually happened (although not in software preservation as far as I'm aware). And, in the rare cases where it's not thrown out by an irritated judge for wasting everyone's time, the judgement is a remedy ("stop doing that") and symbolic damages of $1. And, no, the defendant doesn't have to spend anywhere near what the plaintiff does to defend. You can get to default judgement by doing as little as just filing a few forms and having someone appear once. However, IF the company still existed - the far more likely scenario is they'd just ask OP to remove his blog post and that would be that.
The reason this point matters is that vintage software preservation is historically important yet unfortunately some of our IP laws were created with no thought for what happens in the 'abandonware' scenario where the product is long past economic viability (or even being able to be run) and the company is either gone or no longer cares.
It is always theoretically possible that someone objects to archiving and preservation but the consequences for an individual preserving the software for history and no financial gain, who took reasonable steps to find the copyright owner - are almost always nothing more than a C&D letter - and that's the rare worst case. This is based on decades of experience preserving copyrighted commercial abandonware by non-profit projects like Archive.org and MAME.
My concern is if enough people continually drop the pedantically correct drive-by warning "But you could be SUED" every time some volunteer preservationist saves another title from extinction, people who don't understand there's virtually zero chance in a case like this that OP suffers any meaningful harm simply won't do volunteer preservation or help those who are. It's a form of concern trolling which doesn't help while potentially causing a good thing to happen less.
You can break laws. You just have to realistically estimate the risk. The chance is low, but the impact is high. Maybe you set up an LLC just in case, or you act from the shadows like "Anna"
The government's traditionally very weaponized. I'd err on the side of overestimating risk unless you're certain about the legalities. Your life could be ruined. Most things that are worth doing could ruin your life. You could die in a car crash too. You have to choose which risks to take.
Wow. I kind of regret the effort supplying supporting facts and reasoning demonstrating the impact is almost invariably zero or negligible.
> Most things that are worth doing could ruin your life. You could die in a car crash too.
So... you agree?
> You have to choose which risks to take.
And you have to choose which concerns to troll. Maybe consider the potential unintended harm you may cause of stopping something worth doing before evoking the specter of a scary-sounding consequence. A consequence which is, in reality, very nearly always harmless (in the non-profit, preservation with reasonable prior search scenario) as well as vanishingly unlikely to happen at all.
Finish the bottle if Marcus claims LLMs are 'unreliable stochastic engines' while ignoring that the real bottleneck isn't the model's logic, but the massive security risk of giving them actual system agency.
He’s not entirely wrong about the risks, though. I’ve been trying to set up more 'agentic' workflows recently and it’s a constant battle between convenience and not wanting to hand over my digital keys to a third-party server.
I’ve been experimenting with PAIO (Personal AI Operator) as a middle ground. It’s the first time I’ve seen a 'Bring Your Own Key' (BYOK) architecture that actually feels like a one-click integration rather than a security compromise. It solves that specific Marcus-critique of 'AI being unsafe for real tasks' by keeping the security layer separate from the LLM’s hallucination-prone logic.
Has anyone else here tried their implementation yet? I'm curious if the 'one-click' ease holds up for more complex custom integrations, or if we're still stuck in the 'manual hardening' era for anything serious.
The weird thing is I generally agree with Gary on most of his points. He makes interesting observations that often turn out to be correct. But the Gary Marcus blog post pattern cannot be denied.
Yesterday I tried to explain to my spouse what Moltbook was and why it's interesting - and largely failed. After reading this article, I forwarded it to her. Mission accomplished.
Some nice turns of phrase too...
> The hyper-reality of the platform lies in its circularity. These agents are made of us—trained on our words, rewarded for our preferences, shaped by our data. When they build a society, they build it from the only materials they have: the patterns of human culture, compressed into weights and biases, deployed without the context that gave those patterns meaning.
>The result is not Skynet. It’s Algorithmic Kitsch—a civilization constructed entirely from the discarded cultural artifacts of the human internet, reassembled by machines who know the syntax of society but not the semantics of life. They are dancing to music they cannot hear, in a room we are not allowed to enter.
> "The human brain is mutable, the human "soul" is a concept thats not proven yet and likely isn't real."
The soul is "a concept that's not proven yet." It's unproven because there's no convincing evidence for the proposition. By definition, in the absence of convincing evidence, the null hypothesis of any proposition is presumed to be more likely. The presumed likelihood of the null hypothesis is not a positive assertion which creates a burden of proof. It's the presumed default state of all possible propositions - even those yet to be imagined.
I'd love to buy a newer, better streaming box than the several I already have but after spending quite a bit of time investigating, the state of Android TV-based streaming boxes & sticks is awful. That's one reason NVidia's Shield (Tegra X1 SoC) is still well-regarded despite being a circa 2015 design (the 2019 rev was just a cost reduce/bug fix with the same performance).
Beyond that your choices are to either stick with the same mainstream Google/Amazon/WalMart boxes which are locked down and based on 5+ year-old SoC designs or go with second-tier boxes from Asian vendors on AliExpress/Amazon/eBay, all of which have some different combination of significant compromises:
* Don't work with certain DRM, streaming services or codecs
* Has unreliable manufacturer support (certain firmware works with some DRM/services, next rev fixes one but breaks another)
And even those are built on old hardware designs because there's been no significant advancement in set-top SoC performance for over 5 years. There are only a handful of set-top SoC makers (MediaTek, Amlogic, Rockchip, etc) and while they do occasionally introduce new chips, they mostly only update the video decoding block to support newer codec levels or DRM revisions while keeping the same ancient ARM CPU/GPU cores (or different cores with the same class of 2015-2018 performance).
A good example is the Ugoos AM6B Plus box someone in this thread mentioned as an option for certain use cases. It's been verified to decode DV7 with FEL BUT only works well with local files, not streaming services. And the Amlogic 922x SoC in that box is 5+ year-old tech (I have the same chip in an Amazon Fire TV Stick 4k Max (2nd gen)). The hardware performance of these boxes has been essentially frozen in time due to a 'perfect storm' of factors:
* Most consumers want the cheapest box they can get which plays the main streaming platforms (NetFlix, Amazon, Apple, Disney+, etc). As long as they get a picture of minimally acceptable quality, they don't know or care if the hardware/firmware/drivers properly support the better Dolbyvision levels or adds the enhancement layer or supports ICtCp color space, 12-bit tunneling through RGB or if it handles Source-Based Tone Mapping (SBTM) correctly. They also don't care about playing locally hosted files smoothly or horrendous latency in the Wifi/Ethernet driver stack that nerfs local game streaming.
* DRM is a shit show. The big Hollywood studios require streaming platforms to use specific encryption. So the streaming platform apps will only playback streams on SoCs which have been officially certified (or they nerf the stream to 720p). The certification process is onerous, costly and time-consuming for SoC makers.
* SoC makers, having run the certification gauntlet a couple times now, would like to do it again, approximately... never. On top of that mess, developing and maintaining firmware for their decoding block which properly supports the constantly evolving landscape of divergent codec levels, enhancement layers, color spaces, tone mapping, etc is hard, expensive and requires deep expertise across multiple domains. They just want to sell trays of cheap SoCs and see all the rest as a bottomless money pit eating their slim margins.
NVidia did all this with the Shield and it's grandfathered in on the DRM and they've done a decent job supporting some more recent codecs, levels and layers where they can. But the Tegra X1 platform is 10+ years old now - yet it's still slightly more performant than any other DRM-certified SoC to this day, which just shows what a mess this is.
Which is insanely frustrating if you understand technology platforms, care about actually seeing the full quality modern tech can deliver and would like to do so on a non-ancient hardware platform capable of other trivial things like locally streaming files with actual throughput >100mbps or streaming games with non-glacial latency. But that's just table stakes because the things which could be done with more modern hardware are super-interesting, like AI-based upscaling, frame gen, removing compression artifacts, reformatting content, on-device gaming, etc.
But using standard small form-factor PC/GPU hardware is a non-starter because of DRM certification. So... it would be great if NVidia would make a new Shield based on the new Tegra. But that's a huge new effort and, sadly, NVidia would crazy to divert resources or wafers from the AI-bubble cash printer to anything else - so I highly doubt it's going to happen.
What is it about mobile phone chipsets that makes the unsuitable for a TV stick or STB?
Is it that there is special TV-specific hardware like tuners, HW accelerated audio and video decoders, and PQ/AQ accelerators?
Apple has adapter their A15 chipsets for use in the Apple TV, so it seems possible. But obviously the Apple TV products don't have tuners, aren't driving a display natively, and probably don't have enough I/O interfaces to add all the extra hardware you'd need to embed it in a panel or STB.
It's a good question and I'm not actually sure as I'm not a hardware guy, just a user who's looked into these productz. So far, the popular Android TV set-top boxes (or sticks) I've seen use SoCs that seem dedicated to set-top applications. It may that mobile phone chipsets have different integration to support cellular modems and air interfaces.
There are some boxes which use use Android instead of Android TV but these tend to require using versions of the streaming apps made for mobile phones. I haven't really looked into these as they tend not to work well with remote controls so I haven't been interested.
> I would imagine it's incredibly expensive to maintain.
Everything that flies is expensive to maintain but the costs to maintain most older aircraft tends to be much lower than new ones, sometimes even if certain unavailable parts need to be rebuilt or fabricated. Part of the difference is newer designs tend to use advanced composites and manufacturing techniques which can yield increased performance and efficiency but are expensive and often require specialized techniques to service/replace.
The second factor is that funding, designing, validating and manufacturing new military aircraft platforms has grown astronomically expensive for a huge number of reasons.
Yeah, I full-stopped on that sentence because it was just so bizarre. I can understand making a counter-to-reality claim and then supporting the claim with context and interpretation to build toward a deeper point. But he just asserts something obviously false and moves on with no comment.
Even if he believes that statement is true, it still means he has no ability to model where his reader is coming from (or simply doesn't care).
Dammit! In trying to link to it, I just discovered the wonderful site wisdomofchopra.com is no more. Probably because Chopra recently expanded his grift from books and seminars to an online AI, so the old 2012 random sentence generator site was competition. Oh well. Here's the WisdomofChopra Twitter feed: https://x.com/WisdomOfChopra
Which reminds me, inspired by Deepak Chopra's sayings, actual philosopher Daniel Dennett, who was known for saying things which were not only coherent but often quite insightful, coined the term Deepity: "A statement that is apparently profound, but is actually trivial on one level and meaningless on another."
reply