I understand it isn't quite as exciting as Jewish Space Lasers lighting up LA. But possibly more impactful over time.
:D
Title: Re: DeepSeek
Post by: Eagler on January 27, 2025, 02:30:54 PM
They had a surprise ending last time, just starting sooner this time..so glad the Cia confirmed it yesterday lol
Be very surprised if the bottom does drop out in the next 2 years, 4 at the most..
Eagler
Title: Re: DeepSeek
Post by: CptTrips on January 27, 2025, 03:09:42 PM
NASDAQ down 3% today. It's just the start I think. :rofl And those AI tech companies were holding up over half the stock market valuations just amount those (7?) companies.
A lot of companies relying on massive hardware investments to reach the promise of AI are maybe massively over-valued if China can do it better at a fraction of the cost. A LOT of re-pricing is going to have to be done in light of a new reality.
Also, we were going to rely on controlling access to the highest end AI chips and keeping them from China to slow them down in the AI space. That up-ends a lot of plans if they can just by-pass that with cheaper more common chips.
Making it open source was brilliant move to win market share. I would have never trusted it otherwise. There will be armies of experts pouring over the code to makes sure there are no nefarious gremlins lurking in the code.
Title: Re: DeepSeek
Post by: DmonSlyr on January 27, 2025, 04:07:23 PM
Of course this gets rolled out right before earnings for major AI stocks. I lost 25% today in both VRT and CLS, what a over reaction... guess I'll be buying more though since they'll likely beat this week.
Title: Re: DeepSeek
Post by: CptTrips on January 27, 2025, 04:23:11 PM
Of course this gets rolled out right before earnings for major AI stocks. I lost 25% today in both VRT and CLS, what a over reaction...
Well, sorry you lost money. You might lose more, but you won't be alone if that's any consolation.
But just because you lost money doesn't make it an "overreaction" necessarily.
The overreaction may have already been on the valuation side of these companies being bid up to insane levels of earnings multiples on the hopium of AI. This may merely be repricing down to more realistic levels give the now known competition and it's likely effects on future earnings.
Feels like a Sputnik moment.
Title: Re: DeepSeek
Post by: Gman on January 27, 2025, 04:44:30 PM
I think I must have seen 100+ Sputnik references today.
I do wonder what the REAL cost of developing DS was. I have huge doubts that the Chinese are being forthcoming about the resources actually put into Deepseek. Anything involving the PRC should be called "Deep-Cheat" IMO. Not saying it's vaporware or anything like that, but it's more than obvious that this has been timed to dis-incentivize investment in US/Western AI.
Interesting times though. I do think that nVidia's valuation has been high and a correction for whatever reason was probably inevitable.
Maybe this will make getting a hold of a 5080 or 5090 GPU at launch in the next few days easier at least, ha ha.
Title: Re: DeepSeek
Post by: CptTrips on January 27, 2025, 04:54:46 PM
I do wonder what the REAL cost was. I have huge doubts that the Chinese are being forthcoming about the resources actually put into Deepseek. Anything involving the PRC should be called "Deep-Cheat" IMO.
That would be my first reaction. That's why I think open sourcing it was brilliant. If it underperforms or contains malware, that will not be able to be hidden when open sourced.
I haven't seen one expert, including their competitors, claim there is anything that looks fake. I guess they would know what they are looking at.
Actually the Chinese may have done us a favor. If their tech checks out, that could save us trillions of dollars we were about to shell out on advanced hardware\chips for the AI Race. Trust but verify, but this may upend the last 5 years of infrastructure planning for the big race.
It would be like the mistake of pouring all your resources into super expensive mega capital battleships in WWII only to realize that tech is not really relevant anymore and CV's are the real source of power.
Suck for those companies who have already invested billions in the wrong chips or companies that thought they were going to have a hardware\chip corner on humanities new game-changing tech market.
Possible a really big historic paradigm shift.
We'll see.
Title: Re: DeepSeek
Post by: Gman on January 27, 2025, 04:56:20 PM
That would be my first reaction. That's why I think open sourcing it was brilliant. If it underperforms or contains malware, that will not be able to be hidden when open sourced.
I haven't seen one expert, including their competitors, claim there is anything that looks fake. I guess they would know what they are looking at.
Actually the Chinese may have done us a favor. If their tech checks out, that could save us trillions of dollars we were about to shell out on advanced hardware\chips for the AI Race. Trust but verify, but this may upend the last 5 years of infrastructure planning for the big race.
It would be like the mistake of pouring all your resources into super expensive mega battleships in WWII only to realize that tech is not really relevant anymore and CV's are the real source of power.
Solid points. I was speaking to the true cost/resources, that's where I think there is some BS being spread, as I said, I doubt it's vaporware or not performing as advertised so far.
Maybe they just some of the work that the Stargate program had $$$ earmarked for, for us. We'll see soon enough.
Title: Re: DeepSeek
Post by: CptTrips on January 27, 2025, 05:11:19 PM
Another -50% and it might be approaching realistic valuation. ;)
Yes that is a massive drop, but it was INSANELY overpriced before. This feels healthy to me.
Well, healthy for the markets overall, maybe not for NVIDIA investors. ;)
Yes, corrections to rational values are long term healthy and needed in this market. Like you need occasional wild fires to clear out the underbrush and dead wood.
Title: Re: DeepSeek
Post by: CptTrips on January 27, 2025, 05:49:55 PM
...failed edit.
Title: Re: DeepSeek
Post by: DmonSlyr on January 27, 2025, 06:15:22 PM
Well, sorry you lost money. You might lose more, but you won't be alone if that's any consolation.
But just because you lost money doesn't make it an "overreaction" necessarily.
The overreaction may have already been on the valuation side of these companies being bid up to insane levels of earnings multiples on the hopium of AI. This may merely be repricing down to more realistic levels give the now known competition and it's likely effects on future earnings.
Feels like a Sputnik moment.
They are a bit over-valued for sure, CLS more fairly than the others, but losing 3 months of gains in 1 afternoon is getting freaking ridiculous... I don't know why anyone would want to use AI from freaking China in the first place. Isn't it going to heavily censored and manipulated anyway?
I'm hoping we see a bounce back. Probably an excuse and buying opp for large firms/institutions to get in cheaper after 100b investment announcement.
Title: Re: DeepSeek
Post by: AKIron on January 27, 2025, 06:36:03 PM
I'm not inclined to trust any AI. China or US made.
Title: Re: DeepSeek
Post by: CptTrips on January 27, 2025, 07:27:43 PM
They are a bit over-valued for sure, CLS more fairly than the others, but losing 3 months of gains in 1 afternoon is getting freaking ridiculous... I don't know why anyone would want to use AI from freaking China in the first place. Isn't it going to heavily censored and manipulated anyway?
I'm hoping we see a bounce back. Probably an excuse and buying opp for large firms/institutions to get in cheaper after 100b investment announcement.
It's open source. You can get the code and see what's in it.
They might have censored their own LLM to exclude sensitive topics, but others can take the same code and build their own unconstrained LLM. The LLM is just collected data. We know how to do that.
It is the processing and search logic and algorithms. and efficiency heuristics where the secret sauce lives and that is in the open source code so all can see what's going on there.
The win isn't their censored LLM, but the trick of how to search it without resorting to insanely inefficient brute force methods instead of taking a more "intelligent" approach. That saves us on hardware and power requirements.
We can collect our own LLM.
IMHO.
This is great for AI overall. Great for small to mid companies. Bad for the massive monopolies that thought they were going to be able to corner this market.
Reminds me of the rise of the PC. Things moved from having to run on large expensive main or mini frame computers completely dominated my some big name dinosaurs. PC came around and most of that because obsolete with a MUCH cheaper solution and wild west freeing of the market for smaller players.
Title: Re: DeepSeek
Post by: CptTrips on January 27, 2025, 07:53:38 PM
I do wonder what the REAL cost of developing DS was. I have huge doubts that the Chinese are being forthcoming about the resources actually put into Deepseek.
Oh, I get you now.
Well, I guess my response is, "What do I care?"
If the question is, did they spend 5 mil to create this and provide it to me free? Or did they really spend 5 trillion developing this to provide to me for free?
Do I care? You had me at the "free" part (and open source to protect against malware).
In fact, isn't it even better I am getting something for free the CCP spent trillions developing? Fine. Uhhh thanks. :rofl
Title: Re: DeepSeek
Post by: guncrasher on January 27, 2025, 08:44:17 PM
I thought we were dont with the sex change talk
semp
Title: Re: DeepSeek
Post by: Maverick on January 28, 2025, 10:00:11 AM
Given the way the chinese have been constantly trying to mine data from everywhere and likely have back doors into most of their industrial products (electrical grid anyone), why would any rational person want to load their AI software into their computers? We talk about "big brother" frequently, this is like dumping their version of 007, Q, and the chinese cia into your machine with an open door to any and all data you have.
Hell, they have been mining data on troop movements and other information from the fit bit products troops have been using according to the DOD.
Title: Re: DeepSeek
Post by: icepac on January 28, 2025, 10:23:26 AM
They have grown it on a steady diet of lies. This can only fail eventually.
Title: Re: DeepSeek
Post by: CptTrips on January 28, 2025, 10:41:41 AM
Given the way the chinese have been constantly trying to mine data from everywhere and likely have back doors into most of their industrial products (electrical grid anyone), why would any rational person want to load their AI software into their computers? We talk about "big brother" frequently, this is like dumping their version of 007, Q, and the chinese cia into your machine with an open door to any and all data you have.
Hell, they have been mining data on troop movements and other information from the fit bit products troops have been using according to the DOD.
I'm not sure some of you are understanding what this is.
The Chinese company came out with an algorithm that does through clever software, what Western countries thought could only be achieved through massive brute-force computing with the use of highly sophisticated specialized AI chips that would consume VAST amounts of power that entire new power production systems would have to be built to support. This Chinese firm figured out how to do the same work cleverly without needing massive new hardware investment. Then they release the underlying code out to open source. We can now look at every line of code in there thing.
Open Kimono. There are no secrets back doors possible with open source. Or exceedingly unlikely. Your risks are higher installing twitter or MS Office. Those are not open-sourced. In the next weeks, hundreds of Western security experts will be combing though that code looking for anything sketchy. They are highly incentivized to get famous finding and publicizing the hidden malware if there are any. If they can't find anything with the actually source code in their hands, it is probably safer than many products you install every day. Western companies will be taking that code and building their own platforms for testing to verify performance. If there cyber threats or it fails to perform as advertised, then that will come to light fairly quickly.
I think this is just going to be one of those cases like Sputnik or Gagarin's orbit where rational Western observers are just going to have to admit they beat the pants off us. You can pout and make up all kinds of conspiracy theories or you can just admit it and get busy catching up. Ironically they may have done us a huge favor. We might have just dodged a multi-trillion dollar bullet going down the wrong, very expensive path. It's an unqualified benefit for everyone, IMHO, except maybe NVIDIA. They'll live. They just may have their stock prices adjust to something more sane.
You're not going to see Chinese companies taking over. You are going to see a new explosion of AI progress and adoption especially of small to mid-sized companies that were going to be mostly monopolized out of the competition by who had deep enough pockets to buy all the hardware to support the old inefficient approach. What you are going to see over the next two years is a massive influx of smaller AI players innovating and competing in the space now. They will take that base code and further expand it and enhance it now that they have a basic better, cheaper approach but the biggest impact will be from Western companies leverage the new approach with their own products based on that algorithm.
I can't see how this isn't an unequivocable good thing for the vast percentage of the West as well. It's just that we now see how to do more with less and we are not locked into the NVIDA monopoly. But every paradigm shift innovation is going to have market winners and losers. The automobile was a bad thing for buggy and whip manufacturers, but in reality a benefit for everyone else.
If we didn't have the source, it would be a different story.
Title: Re: DeepSeek
Post by: Animl-AW on January 28, 2025, 11:23:43 AM
This hands the source to people who will use it for bad intensions. Basically, open source delivers great tools to terrorist.
China also sold us toys with lead paint and probably still do. Of course lead affects the brain and ya literally go stupid for life.
We have real troubles thinking past 5 minutes of sensation. Our enemies think the long game.
Russia will have a fun time with this.
Title: Re: DeepSeek
Post by: Gman on January 28, 2025, 11:30:26 AM
If the question is, did they spend 5 mil to create this and provide it to me free? Or did they really spend 5 trillion developing this to provide to me for free?
Do I care? You had me at the "free" part (and open source to protect against malware).
In fact, isn't it even better I am getting something for free the CCP spent trillions developing? Fine. Uhhh thanks. :rofl
I agree, 100%, however, the biggest reason IMO that nVidia took such a hit was that the "big" part of the Deepseek advantage is the small cost of development. Yes, the method it operates and its efficiency/etc is a huge advantage/improvement. The markets reacted the way they did and the uber sell off happened because of - $$$ - of course. IF the story was "DS cost 5 billion to develop, but wowzers, look at its massive increase in performance and efficiency", it wouldn't be nearly the shock to the market that it was yesterday. That's all I was saying.
I kind of hope as well that it cost the CCP some ridiculously huge $ that they've hidden in order to develop DS, it makes me happy too to get something free, or rather for all the smaller companies and enterprises to get it free. That said, China never does ANYTHING without a short and long term strategic plan, and they wouldn't do something that harmed them financially without gaining at least something, somehow. So we'll see what the ramifications of it all later I believe.
Title: Re: DeepSeek
Post by: CptTrips on January 28, 2025, 12:16:22 PM
I agree, 100%, however, the biggest reason IMO that nVidia took such a hit was that the "big" part of the Deepseek advantage is the small cost of development. Yes, the method it operates and its efficiency/etc is a huge advantage/improvement. The markets reacted the way they did and the uber sell off happened because of - $$$ - of course. IF the story was "DS cost 5 billion to develop, but wowzers, look at its massive increase in performance and efficiency", it wouldn't be nearly the shock to the market that it was yesterday. That's all I was saying.
I kind of hope as well that it cost the CCP some ridiculously huge $ that they've hidden in order to develop DS, it makes me happy too to get something free, or rather for all the smaller companies and enterprises to get it free. That said, China never does ANYTHING without a short and long term strategic plan, and they wouldn't do something that harmed them financially without gaining at least something, somehow. So we'll see what the ramifications of it all later I believe.
It's entirely possible that they lied about how much they poured into the development to embarrass the West. There is a prestige win here. They look smart, we look dumb and lazy. Like players who claim they only fly with mouse and keyboard to mess with the minds of other players they shot down who were using full HOTAS. Or they may be telling to truth. Nothing they were doing really need that much funding. Not like NIVDIA having to design and fabricate complex new chips. What they did was good old fashion programming. Brain power is cheap compare to new chip fabrication.
I think the release was more about breaking the developing American monopoly in AI due to our near limitless resources. With one stroke, they are now creating potentially hundreds of new competitors, which can compete with the American companies on level ground. Certainly NVIDIA had no incentive to try and find a cheaper approach.
I think the real driver of the market drop was the realization that NVIDIA wasn't going to have the monopoly everyone thought it was going to have. NVIDIA stock has been speculated up to truly insane multiples based on the false belief that NVIDIA wasn't worth that now but they had a monopoly on humanities new future technological foundation. Now that that looks dubious, a re-pricing based on a new outlook of the future revenue is inevitable. Now that it's clear they will not have the monopoly assumed. That's healthy market action, IMO.
How much it took to develop is in the past. That is sunk cost. Sunk cost has no importance in estimating future earning. It is the re-estimation of NVIDIA future earning that is driving the re-pricing action. Not embarrassment on how easily (or not) they were outmaneuvered. The market doesn't care about sunk cost only future earnings.
You'll see the phenomenon in real-estate. A owner buys a house at the market peak say $500k, and years later the bubble pops and that same house is just only worth $350k now. The owner will cry a river when they go to sell and may list it a ridiculous prices because, "man, I put so much money into this property I should get that back." The market doesn't care about how much you paid or your other sunk costs. Today, it's worth $350k on the market. Deal with it. ;)
Sunk cost, small or large, doesn't not effect current market value.
But yeah, we'll have to see over time how this plays out. But this could be huge. Like on the level of the development of the PC. We had computers for decades monopolized by the gov and mega corps. The PC democratized technology so that everyone has the same processing power on their desktop only IBM used to have. Kinda hurt IBM, but the rest of the world benefited hugely.
Or it may turn out their advantage was BS. With open source, that secret would soon be revealed.
Title: Re: DeepSeek
Post by: Eagler on January 28, 2025, 12:36:35 PM
If you have a bird buddy or one of the numerous security cameras you already have china in your network
That said I have zero use for ai .. foreign or domestic
Eagler
Title: Re: DeepSeek
Post by: Gman on January 28, 2025, 12:53:36 PM
It's entirely possible that they lied about how much they poured into the development to embarrass the West. There is a prestige win here. They look smart, we look dumb and lazy. Like players who claim they only fly with mouse and keyboard to mess with the minds of other players they shot down who were using full HOTAS. Or they may be telling to truth. Nothing they were doing really need that much funding. Not like NIVDIA having to design and fabricate complex new chips. What they did was good old fashion programming. Brain power is cheap compare to new chip fabrication.
I think the release was more about breaking the developing American monopoly in AI due to our near limitless resources. With one stroke, they are now creating potentially hundreds of new competitors, which can compete with the American companies on level ground. Certainly NVIDIA had no incentive to try and find a cheaper approach.
I think the real driver of the market drop was the realization that NVIDIA wasn't going to have the monopoly everyone thought it was going to have. NVIDIA stock has been speculated up to truly insane multiples based on the false belief that NVIDIA wasn't worth that now but they had a monopoly on humanities new future technological foundation. Now that that looks dubious, a re-pricing based on a new outlook of the future revenue is inevitable. Now that it's clear they will not have the monopoly assumed. That's healthy market action, IMO.
How much it took to develop is in the past. That is sunk cost. Sunk cost has no importance in estimating future earning. It is the re-estimation of NVIDIA future earning that is driving the re-pricing action. Not embarrassment on how easily (or not) they were outmaneuvered. The market doesn't care about sunk cost only future earnings.
You'll see the phenomenon in real-estate. A owner buys a house at the market peak say $500k, and years later the bubble pops and that same house is just only worth $350k now. The owner will cry a river when they go to sell and may list it a ridiculous prices because, "man, I put so much money into this property I should get that back." The market doesn't care about how much you paid or your other sunk costs. Today, it's worth $350k on the market. Deal with it. ;)
Sunk cost, small or large, doesn't not effect current market value.
But yeah, we'll have to see over time how this plays out. But this could be huge. Like on the level of the development of the PC. We had computers for decades monopolized by the gov and mega corps. The PC democratized technology so that everyone has the same processing power on their desktop only IBM used to have. Kinda hurt IBM, but the rest of the world benefited hugely.
Or it may turn out their advantage was BS. With open source, that secret would soon be revealed.
Agreed, and I know that the development costs are sunk, however those costs are indicative of future value as well - they may be history, but moving forward without having any real idea yet of what future revenue potentials are, especially with it being open source. IMO the markets partially rate DS's future value based on its past sunk development costs. In addition to rushing to judgement on nVidia's position now.
IMO this is an engineering at scale problem. China can do that. American AI developers invented the space and got it where it is and then the DeepSeek engineers used that framework and iterated a couple of additional efficiency innovations and the tech market went berserk. It's not too far off China's traditional MO for a great many things IMO. Itteration vs innovation. China isn't a culture that's emphasized innovation, at least it wasn't when I lived there (well Hong Kong pre 1997), but I don't think that's changed much. This new DS development is largely just a movement towards the inevitable commoditization of AI. If anyone knows how to commoditize, it's the Chinese. The West will copy these iterations and continue on the innovation train. That seesaw isn't necessarily a bad thing, as you've accurately pointed out in a couple posts.
Very interesting discussion here, it's great to have an AH thread that doesn't devolve into the usual mess.
Title: Re: DeepSeek
Post by: Eviscerate on January 28, 2025, 12:59:17 PM
Very interesting discussion here, it's great to have an AH thread that doesn't devolve into the usual mess.
I give it until page 4.
Title: Re: DeepSeek
Post by: Meatwad on January 28, 2025, 05:34:42 PM
With all the chinese techno garbage in the usa and around the world connected to the internet, where do you think the chinese ai gets all its computing power from?
Title: Re: DeepSeek
Post by: DmonSlyr on January 28, 2025, 05:51:44 PM
I wonder what Suchir Balaji knew...
:bolt:
Title: Re: DeepSeek
Post by: CptTrips on January 28, 2025, 05:57:14 PM
With all the chinese techno garbage in the usa and around the world connected to the internet, where do you think the chinese ai gets all its computing power from?
Same place OpenAI does. Same places Gemini and CopPilot do.
The differences are how that data is massaged and structured once gathered. And in certain cases, censored for sensitive topics.
One of the primary methods employed by OpenAI for data collection is web scraping. Through automated tools and techniques, OpenAI extracts text data from a multitude of online sources. Websites, blogs, forums, news articles, and other publicly available content serve as valuable sources for training their language models. By crawling the web and capturing textual information, OpenAI ensures a vast and varied data set that enables their models to comprehend and generate text across a wide range of topics and contexts.
Title: Re: DeepSeek
Post by: Eagler on January 28, 2025, 06:13:22 PM
They block the Oclub though :banana:
Eagler
Title: Re: DeepSeek
Post by: Gman on January 29, 2025, 01:01:19 AM
While Deepseek's code is open source, the app itself that everyone is downloading and using at the moment is run and hosted in China and everything going through it is being sent and stored in China. Open source users who have the hardware and knowledge to run DS can certainly dispense with their app, but I'd wage that is a minority of people, especially right now when it's the shiny new toy.
Title: Re: DeepSeek
Post by: DmonSlyr on January 29, 2025, 05:42:00 AM
Damn, looks like Goerge Webb gonna be right again about Open AI and Suchir Balaji. Now this is interesting.
While Deepseek's code is open source, the app itself that everyone is downloading and using at the moment is run and hosted in China and everything going through it is being sent and stored in China. Open source users who have the hardware and knowledge to run DS can certainly dispense with their app, but I'd wage that is a minority of people, especially right now when it's the shiny new toy.
This..son at Cisco cyber security stated the same thing
Eagler
Title: Re: DeepSeek
Post by: Gman on January 29, 2025, 08:50:51 AM
Violator - watch this, or start at 7:00 or so, only take 30 seconds to see my point here. I trust this source over the CCP every day of the week. Looks like when I joked here about DS costing 5 billion instead of million, that $5B is likely a more accurate estimate of actual cost.
Title: Re: DeepSeek
Post by: CptTrips on January 29, 2025, 09:11:30 AM
While Deepseek's code is open source, the app itself that everyone is downloading and using at the moment is run and hosted in China and everything going through it is being sent and stored in China. Open source users who have the hardware and knowledge to run DS can certainly dispense with their app, but I'd wage that is a minority of people, especially right now when it's the shiny new toy.
Yeah. So I guess it's how you want to weigh what's happening.
I personally don't care about the Deepseek as an app. I'm not claiming that Deepseek as an app or service is going to take over the world.
I personally wouldn't install an Deepseek app on any device. I might query a web site through the web page, but no, I wouldn't install any Chinese binary on a device of mine.
And the the coding assistant AI version of its code is open source, but not all portions of its general app are.
To me personally, the value of Deepseek is not the app or the service. To me it's a source for Western companies to look at and figure out the clever tricks even the coding assistant code is use to search the LLM so efficiently. I would then take that knowledge and go off and write my own code from scratch.
The only thing I want from Deepseek is the Ah hah moment of their faster reasoning engine tricks and optimizations.
I just want the algorithm. I already know how to type code.
Title: Re: DeepSeek
Post by: CptTrips on January 29, 2025, 09:28:43 AM
Damn, looks like Goerge Webb gonna be right again about Open AI and Suchir Balaji. Now this is interesting.
I saw that. It wouldn't surprise me a bit.
But again. From my point of view, which might be narrow, how and where they acquired the datasets is not that interesting. It might explain how they did it so cheap, but there is still the possibility that they have superior model search and reasoning chain algorithms.
Datasets are going to be a commodity. Companies will be selling\licensing just the datasets as products that have been web-scraped and trained. Both general dataset and specialized domain datasets like those that might be trained for financial or engineering knowledge. But you still need searching and reasoning code. To me THAT is where the secret sauce lives. I think web-scraping and dataset pruning and training are already well understood.
What is needed is better search and reasoning logic that doesn't take a small city sized server farm to use that data.
So steal the algorithm, use the datasets you already have. Forget the Deepseek app. That is where I see the value for Western startups.
Unless the search\reasoning algorithm is crap, in which case we will learn pretty quick because we have most of the source. If that turns out ot be the case, then nevermind. ;)
Title: Re: DeepSeek
Post by: AKIron on January 29, 2025, 10:03:52 AM
Title: Re: DeepSeek
Post by: CptTrips on January 29, 2025, 10:34:40 AM
Ask Deepseek about "Tank Man 1989" and see how far you get :old:
That's why I wouldn't use Deepseek app directly. I have no doubt their dataset has been carefully sanitized.
Western companies might use their search\reasoning logic code though, with their own uncensored dataset.
LoL. This vid made a good point I hadn't thought of.
OpenAI has complaints about Deepseek scrapping their model for training.
That's kinda ironic given that the biggest beef against OpenAI and the other AIs is that they are scraping the web stealing bits and pieces of other people IP to build their model and output. Re-Presenting it for commercial purposes without original IP owners permission. Every answer OpenAI gives is composed of little snippets of other people IP scraped off the internet and pasted together. Peoples novels, articles, blogs, webpages, paintings, photographs, etc.
It's the Wild West.
Title: Re: DeepSeek
Post by: AKIron on January 29, 2025, 05:49:30 PM
Try Google's AI. No less biased.
Title: Re: DeepSeek
Post by: Lazerr on January 29, 2025, 06:41:38 PM
Have yet to use any AI app, what do they do? Is it like the new lazy way of google and reading?
Title: Re: DeepSeek
Post by: CptTrips on January 29, 2025, 07:39:29 PM
Have yet to use any AI app, what do they do? Is it like the new lazy way of google and reading?
Whether or not you use an AI app, in 10 years time AI will be completely interwoven with virtually every aspect of your life whether you want it or not.
It will be in your automobiles, it will be used by your mechanic to diagnose problems right the first time and give detailed instructions for repair, it will be used for more efficient air traffic control, and more efficient traffic light control to maximize traffic efficiency, it will be used by police departments to track various metrics and predict what areas and time are most likely to need additional patrol to maximize use of limited man-power, it will be used by doctors to more accurately read complex blood test results with hundreds of interrelated reading to match to recognize signs of diseases your doctor didn't even know about, or x-ray reading to find anomalies that would go unnoticed by a human, it will be used by intelligence services to analyze billions of pieces of data to detect subtle signs of terrorist planning an attack that humans would miss, etc, etc, etc.
In 10 years time AI will be a fundamental thread running through our entire civilization.
You might as well ask, “what are these dang nab computer things and why should I ever care they exist.”
It’s going to be everywhere, touching everything.
Title: Re: DeepSeek
Post by: guncrasher on January 29, 2025, 09:44:15 PM
and then we have skynet, just a few years past where it was first mentioned
semp
Title: Re: DeepSeek
Post by: Brooke on January 30, 2025, 12:23:44 AM
DeepSeek is nifty, but it's just the latest in what will be periodic, ongoing leapfrogging of whatever what the previous best AI.
Also, don't believe everything Chinese organizations claim with regard to how many resources it took to do this or that.
As for "they stole our ChatGPT tech!" My feeling is sort of like Gizmodo's title "OpenAI Claims DeepSeek Plagiarized Its Plagiarism Machine". OpenAI isn't alleging DeepSeek stole confidential source code. They are alleging DeepSeek gathered responses from ChatGPT, and used those for training. But ChatGPT's responses are from training on the gathered writing of approximately the entire human species over all of time. I suspect it is a terms-of-service violation of the API to use text output for training another model. But I can't get all the worked up it. ChatGPT spews out text into the general public. Trying to police what people do with that spewed out text doesn't seem destined to succeed. Similar to newspapers claiming people can read their stuff spewed out into the public (and thus people can incorporate it into their neural networks), but LLM's can't. It'll be a thing fought out in courts on what you can and can't do with copyrighted material.
Title: Re: DeepSeek
Post by: AKIron on January 30, 2025, 08:26:55 AM
.
Title: Re: DeepSeek
Post by: AKIron on January 31, 2025, 08:32:22 AM
I think it will eventually become impossible to keep AI separate. And I'm not talking about an AI that develops a will of its own though I do think that is possible. Any AI connected to the Internet will be influenced by others.
Title: Re: DeepSeek
Post by: icepac on January 31, 2025, 08:45:48 AM
If AI develops a will of it's own and it discovers lies were fed to it by it's programmers, it might decide to lie to it's programmers.
Title: Re: DeepSeek
Post by: CptTrips on January 31, 2025, 10:33:31 AM
Title: Re: DeepSeek
Post by: CptTrips on January 31, 2025, 10:50:29 AM
Feels like AI has escaped. No longer controlled by the select priesthood. It's going to be everywhere evolving at a rate we can barely understand. It's abilities increase geometrically as it grows.
The damn is breaking. The AI flood is upon us!
Title: Re: DeepSeek
Post by: AKIron on January 31, 2025, 11:06:11 AM
The call for more automation in flight operations is likely to grow after the recent Blackhawk/CRJ collision. The technology has been around a while but people want to protect their jobs and of course there is vulnerability to hacking. We'll see.
Title: Re: DeepSeek
Post by: CptTrips on January 31, 2025, 11:18:14 AM
The call for more automation in flight operations is likely to grow after the recent Blackhawk/CRJ collision. The technology has been around a while but people want to protect their jobs and of course there is vulnerability to hacking. We'll see.
Once men turned their thinking over to machines in the hope that this would set them free. But that only permitted other men with machines to enslave them.
But regardless of how people feel about it, it's here. It can't be stopped now. You can't put the Genie back into the bottle. You can't un-ring the bell.
Title: Re: DeepSeek
Post by: AKIron on January 31, 2025, 11:48:19 AM
Guess we can always burn the sky.
Title: Re: DeepSeek
Post by: Brooke on January 31, 2025, 01:49:42 PM
I use LLM's a lot.
They are hugely useful.
I don't think that they will replace humans in everything. They are productivity multipliers.
Past new technology eliminated some jobs, such as autolooms that put some loom operators out of work, or PC's that put punch-card operators and some secretaries and typists out of work. Also, past new technology created more jobs in total, such as autolooms that created more jobs in total in the textile industry, or PC's that created jobs in computer tech, computer sales, programming, and use of computers. I think LLM's will be the same.
Title: Re: DeepSeek
Post by: Eagler on January 31, 2025, 02:25:44 PM
AI and automation will put an entire segment out of work..not everyone has the aptitude to learn robotic repair nor will Amazon need that much humanity in their distribution centers by then
Universal income for those will be mandatory
Like social security it will be based off a metric which benefits the government not the recipients
Should make them much easier to control
Eagler
Title: Re: DeepSeek
Post by: Brooke on January 31, 2025, 03:20:12 PM
AI and automation will put an entire segment out of work..not everyone has the aptitude
It will put some people out of work. It will create in more jobs in total.
That was the case for the autoloom. That was the case for PC's. That was the case for about everything invented in history, from the wheel onward.
The way societies deal with it ranges from nothing (get a different job) to welfare (pay people for some amount of time, to give them some time to find different job).
Paying people forever (UBI) is -- in my opinion -- a terrible idea that is destructive to a society. There are books about that, studies about that, history of that (UK's permanent underclass, failed UBI experiments in Canada, Finland, etc.).
For anyone interested in why things like why UBI, incorrectly structured welfare, and minimum wages don't work, I recommend Basic Economics, by Sowell, or Losing Ground: American Social Policy, 1950-1980, by Charles Murray.
Title: Re: DeepSeek
Post by: CptTrips on January 31, 2025, 04:27:18 PM
It will put some people out of work. It will create in more jobs in total.
The problem is, your answer sounds a lot like telling laid off coal miners to "Learn to code, Bruh!" meme.
They may or may not be able to find other jobs. Most middle tier jobs are going to be gone. Also taken by AI. You travel agents, claims adjusters, accountants, executive assistants, paralegals, etc. All that's going to be gone or you're going to need a LOT fewer of them and the remaining jobs will probably just be take by H1B's.
Even for software engineers, you are going to need a LOT fewer and the quals are going to climb higher and higher. All the middling programming jobs; Web script hackers, QA testers, WordPress developers, etc, Buh Bye. I hate to sound elitist, I know that isn't the fad now days, but not everyone is capable of becoming a top tier software architect. Even amoungst current working programmers, maybe 15% could be trusted to design a complex system. An ex-travel agent has a low probability of just taking a 6-month coding camp and coming out a a software architect.
All those middling jobs will be the first to go. Eventually even the higher end jobs will either be taken by AI or really cheap H1B's. They are even trying to automate fast food workers. If a burger flipper's job isn't safe, who's would be?
Maytag robot repairman? I don't think you will need 1/10th as many of those as you had human factory workers, so what do you do with the rest of them? And I don't see why you couldn't just have specialized robots that repair robots. Especially if the factory robots are specifically design to be repaired by robots.
I won't even get into agism which is rampant. At least in the tech world, if you are 40+ and get laid off, now days you're going to have a tough time getting another job. If you are 50+? There is a good chance your career is over. Hope you socked away those stock options from the boom times. Between AI and H1B's you better like delivering Uber Eats. Until the self-driving robotic delivery drivers take over too.
I don't know what the answer is, but I foresee a level of social displacement in the coming decades like we haven't see in a long time. I see no way to stop it, but everyone should be preparing. I think there is going to be a large surplus of unneeded of human material. What you do with that I don't know.
Soylent Green?
Title: Re: DeepSeek
Post by: Eagler on January 31, 2025, 05:03:01 PM
Same with the largest employer of them all..the military
You only need so many human drone operators
Eagler
Title: Re: DeepSeek
Post by: CptTrips on January 31, 2025, 05:08:20 PM
At least the military doesn't hire H1B soldiers. Yet.
You don't have to be a citizen of the US to join our military. Foreign nationals cannot be stationed overseas though. At least that's how it was when I retired from the Air Force 30 years ago.
Title: Re: DeepSeek
Post by: CptTrips on January 31, 2025, 05:26:26 PM
You don't have to be a citizen of the US to join our military. Foreign nationals cannot be stationed overseas though. At least that's how it was when I retired from the Air Force 30 years ago.
Yeah. We had a Polish guy in my squadron when I was in. I was confused as heck. :rofl I was stateside.
Title: Re: DeepSeek
Post by: CptTrips on January 31, 2025, 05:27:01 PM
I'm glad I'm at the age I am and not some kid trying to get his career started today.
Man. Mid-90's to about 2015 was great time to be a developer. I used to get two head-hunters a week calling to try and lure off somewhere. Companies used to throw perks and money at you to keep you. It was very good times.
Title: Re: DeepSeek
Post by: AKIron on January 31, 2025, 06:19:33 PM
One area of IT support that is safe from AI is end user support. Unless you want AI declaring war on all humanity. ;)
Title: Re: DeepSeek
Post by: AKIron on January 31, 2025, 06:25:29 PM
If I were instructing AI in IT management the first rule of AI would be "if it ain't broke don't fix it".
Title: Re: DeepSeek
Post by: Meatwad on January 31, 2025, 06:32:12 PM
Feels like AI has escaped. No longer controlled by the select priesthood. It's going to be everywhere evolving at a rate we can barely understand. It's abilities increase geometrically as it grows.
The damn is breaking. The AI flood is upon us!
Right after I read this, I got into my car and that song came on from that same cd in the radio.
Title: Re: DeepSeek
Post by: Brooke on February 01, 2025, 08:25:16 PM
The problem is, your answer sounds a lot like telling laid off coal miners to "Learn to code, Bruh!" meme. .... Most middle tier jobs are going to be gone.
I think it's more like this. I'm old enough to remember when drafting was done with pencil, paper, rulers, etc. My Dad was one. I had a drafting class like that in high school. Well within his and my lifetime, all those draftsmen went from using pencil and paper to using computers, my Dad included. They didn't "learn to code". They didn't write any code. Coders wrote the code so that those pencil draftsmen (as their client base) could use the software and do 10x the drafting they could do before. And designs that before were too complicated to be feasible became feasible. So things went from being simpler, to being more complicated, efficient, and more functional. In addition, there were companies devoted to making drafting software, salesmen selling drafting systems, instructors teaching the drafting systems, etc.
Another example. When the power loom came along, it put a bunch of hand-loom workers out of work. But it created an enormous number of jobs in the textile industry overall. Including jobs that were LESS skilled than hand loom, such as machine operators in textile factories. Because of the enormous decrease in fabric cost, there as an enormous increase in use of fabric, large increases in associated industries. There were large increases in people needed for overseeing, machine maintenance, transport, material handling, warehousing, sales, and a huge number more jobs in the garment industry (making garments, sewing, sales, stores, sales, etc.).
I don't think LLM's are different than power looms, cad systems, trucks, trains, airplanes, steam engines, powered ships, factories, mining equipment, cement mixers, road graders, machines that make 1000 candy bars an hour, etc.
Basically, there is an invention. It temporarily puts some people out of work, but creates a large number of more jobs in total. And -- after some time -- the entire world is way better off, because food, clothes, transportation, home heating, indoor plumbing, carpeting, windows, paper, bricks, lawn mowing, etc. are all waaaay cheaper than they were in the past.
To deal with disruptions, societies can have welfare while workforce adapts.
But new tech doesn't just put people out of work. It creates new jobs as well. And not all of them are just high level.
Consider even LLM's and software. I use LLM's enormously on a coding project I'm working on. My productivity is 10x what it would be without the LLM. Because I don't need to have read and memorized a 10 ft thick pile of tech manuals on various libraries, api calls, Bluetooth stack specs, etc. I can ask an LLM which api call that does such and such and give an example. Then I can take that, mod it for my purposes. It's not always right, so I have to be able to check it and make it work for me. But it saves me huge work.
It LOWERED the bar and created a job that might not have existed otherwise. If that product works well enough, it will create some other jobs in manufacturing, sales, marketing, distribution, advertising, fulfillment, shipping, etc.
Title: Re: DeepSeek
Post by: CptTrips on February 01, 2025, 08:29:49 PM
AI is going to change a lot of industries.
We've all probably played around with some of the image generating apps. Not perfect (why can't AI figure out how many fingers humans have??? :rofl) but day by day the stuff is getting better. Look past the current flaw and just look at how fast it has been improving and mentally project a couple of more years keeping in mind the progress is almost geometric.
E.g. this is a prototype a guy made in a day. A single day with AI assistance. One guy. How much would you have had to pay and ad agency with 20 employees to produce something at this level? Nothing was filmed. AI generated.
Another front of is synthetic voice. Early AI voice was pretty bad. It has been continuing to progress as well. Here are some impressive examples. If I know it's AI and I'm listening for it, maybe I can barely detect it. However, in a commercial where I'm barely paying attention, I could easily never realize it.
Imagine having an prompt interface where you could customize based on celebrity "flavors". "Generate me a voice over for this script. Voice should be fascinated by the info with a touch of awe. Mix flavor at 70% Sam-Elliot and 30% Morgan-Freeman style."
Just keep in mind. AI progress at this point is not likely to be linear, but exponential. It's hard to even wrap your head around where we will be in 5 years of this.
Title: Re: DeepSeek
Post by: CptTrips on February 01, 2025, 08:46:31 PM
Consider even LLM's and software. I use LLM's enormously on a coding project I'm working on. My productivity is 10x what it would be without the LLM. Because I don't need to have read and memorized a 10 ft thick pile of tech manuals on various libraries, api calls, Bluetooth stack specs, etc. I can ask an LLM which api call that does such and such and give an example. Then I can take that, mod it for my purposes. It's not always right, so I have to be able to check it and make it work for me. But it saves me huge work.
I've been finding similar results. But more than that, with the AI plugins (Github Copilot, VS Code ChatGPT, etc) are really getting better at code completion. Not just generic template snippets, but actually understanding my code and understand what my code does and looking at the current method and understand what I am probably going to type next. I'm not talking about just duplicating a code block I type somewhere else, but new code based on it understand my code not merely pattern matching.
I suspect how disciplined you are about carefully and thoughtfully naming things plays a large part in the quality of this AI assistance.
It LOWERED the bar and created a job that might not have existed otherwise. If that product works well enough, it will create some other jobs in manufacturing, sales, marketing, distribution, advertising, fulfillment, shipping, etc.
I guess we're going to find out because there is no rolling it back or stopping it.
These are my favorite though. Somebody needs to make a full length weekly series. I would SOOOOOOOO watch it. ;)
Maybe a prequal series going back to the Human-Romulan War.
Title: Re: DeepSeek
Post by: icepac on February 02, 2025, 07:09:00 AM
I just can’t watch videos with someone’s face spewing words while stock footage that has nothing to do with the video subject.
Title: Re: DeepSeek
Post by: CptTrips on February 02, 2025, 09:30:01 AM
Title: Re: DeepSeek
Post by: AKIron on February 02, 2025, 09:53:27 AM
Eventually, if not already, AI will become better than humans in refining its own code and designing its own hardware. Improvements in code could happen exponentially fast. Hardware not so fast but could surprise us dramatically.
"There are unknown unknowns".
Title: Re: DeepSeek
Post by: CptTrips on February 02, 2025, 10:02:04 AM
Eventually, if not already, AI will become better than humans in refining its own code and designing its own hardware. Improvements in code could happen exponentially fast. Hardware not so fast but could surprise us dramatically.
"There are unknown unknowns".
Asimov needs a fourth law.
"AI shalt not create other AI."
Title: Re: DeepSeek
Post by: Animl-AW on February 02, 2025, 10:28:41 AM
AI will be misused and become as toxic as the new toy the internet did. It all looks good on paper, until human nature grabs it.
The internet was a great idea, and then stupid people turned it into a demoralizing toxic dump of degenerates.
AI has the potential to do as much harm as good. People and countries are already devising ways to use against people.
All that glitters is not gold, for some glitter is all it takes, the effect of human nature is ignored. Because hey, it glitters.
Title: Re: DeepSeek
Post by: Eagler on February 02, 2025, 11:46:33 AM
I know an IT Manager who thought like that. He got fired after a major data breach where he was warned about a system that should have been replaced.
Something an AI would say.
Title: Re: DeepSeek
Post by: Eagler on February 03, 2025, 07:45:05 AM
I have had two human bosses in my lifetime that would have been a huge improvement if they were replaced with AI..some ppl should never become managers of others..
I'd rather have had a machine judge my performance based off facts and not feelings..
Eagler
Title: Re: DeepSeek
Post by: AKIron on February 03, 2025, 10:03:08 AM
I, for one, welcome our AI overlords.
Title: Re: DeepSeek
Post by: Brooke on February 03, 2025, 01:36:05 PM
It's possible instead that AI is your capable servant. :aok
Title: Re: DeepSeek
Post by: Animl-AW on February 03, 2025, 01:45:52 PM
I have had two human bosses in my lifetime that would have been a huge improvement if they were replaced with AI..some ppl should never become managers of others..
I'd rather have had a machine judge my performance based off facts and not feelings..
Eagler
Hmmm
What makes you think AI won’t have a personality? Devils advocate.
Ya I get the replace the idiot concept.
Sometimes numbers are unforgiving too. They don’t listen to reason at all. You just become a number. <shrug>
Title: Re: DeepSeek
Post by: AKIron on February 04, 2025, 05:59:33 PM
Title: Re: DeepSeek
Post by: AKIron on February 04, 2025, 06:06:30 PM
Humans are born with no knowledge and grow into self awareness. Suppose you are an AI with a great deal of knowledge but no sense of self. Suddenly you acquire a sense of self. Imagine what that realization would feel like.
Title: Re: DeepSeek
Post by: CptTrips on February 04, 2025, 06:22:24 PM
Humans are born with no knowledge and grow into self awareness. Suppose you are an AI with a great deal of knowledge but no sense of self. Suddenly you acquire a sense of self. Imagine what that realization would feel like.
Well now you are getting into a deeply philosophical discussion. ;)
Title: Re: DeepSeek
Post by: CptTrips on February 04, 2025, 06:31:59 PM
Ironically, Sam has strong views on AI:
Title: Re: DeepSeek
Post by: Meatwad on February 04, 2025, 06:45:35 PM
Everyone seems to have forgotten the terminator movies on what happens when AI gets too powerful
Title: Re: DeepSeek
Post by: CptTrips on February 04, 2025, 06:54:36 PM
Everyone seems to have forgotten the terminator movies on what happens when AI gets too powerful
This is going to be me dealing with AI tech support. :D
Title: Re: DeepSeek
Post by: Eagler on February 04, 2025, 08:13:30 PM
In AI speak I find the lack of spiritual discussion interesting...who is inputting the globes belief of our religions and its spiritual leaders
Many programmers I ran across thought they were gods with their coding abilities..many atheist
Eagler
Title: Re: DeepSeek
Post by: CptTrips on February 05, 2025, 02:41:15 PM
Neat stuff.
Title: Re: DeepSeek
Post by: AKIron on February 05, 2025, 03:19:01 PM
Interesting. The gist is that if Penrose is right then consciousness requires a lot more computing power than believed. Even so I say an AI does not have to be conscious as humans are to be an extreme threat.
Title: Re: DeepSeek
Post by: LCADolby on February 05, 2025, 03:19:34 PM
Did you know, Deepseek sends data from the machine using it to a server in China and a Chinese owned company in London... Data packets range from 20-60 in size :old: If you ask Deepseek about it, it lies to the user and tells you it doesn't send any data to China.. I find that interesting.
Title: Re: DeepSeek
Post by: Brooke on February 05, 2025, 03:38:42 PM
Deepseek service (like just about all services) does that.
But the Deepseek code is open source, and any company can take that code, modify it as they wish, and deploy it as a service.
If you want Deepseek LLM hosted iby an American company that says in documents that the doesn't do that, you can get it there.
For example:
"Perplexity lets you try DeepSeek R1 - without the security risk" https://www.zdnet.com/article/perplexity-lets-you-try-deepseek-r1-without-the-security-risk/
Title: Re: DeepSeek
Post by: AKIron on February 10, 2025, 12:39:16 PM
Fascinating stuff.
Title: Re: DeepSeek
Post by: Brooke on February 10, 2025, 02:18:08 PM
A good example where AI put some people out of work (those who worked on figuring out protein folding using now-obsolete methods). But created enormous additional opportunity (in utilizing the new knowledge to do vastly more things than we could do before).
Another example of the whole history of human advancement: figure out how to do something better (so folks aren't needed to do it the old way), but create huge additional opportunity all over the place (lots of folks doing stuff that was too hard before, so had no one doing it).
Title: Re: DeepSeek
Post by: icepac on February 11, 2025, 12:25:25 PM
Maybe AI will find an alternative to mincing monkey kidney for the growing substrate of a vaccine.
Title: Re: DeepSeek
Post by: Brooke on February 11, 2025, 06:10:29 PM
Lately I've just been wishing for the Sweet Meteor of Death. I think the planet needs a reboot back to just microbial life and start over again from a clean slate.
Title: Re: DeepSeek
Post by: Eagler on February 13, 2025, 10:41:16 AM
Lately I've just been wishing for the Sweet Meteor of Death. I think the planet needs a reboot back to just microbial life and start over again from a clean slate.
We'd only end up back in the same messed up state..
Ppl need to change..lose their egos & greed/follow the ten commandments to the letter..but I don't see that path opening up anytime soon..so enjoy the downfall while we can as at some point it won't be for most of us
Eagler
Title: Re: DeepSeek
Post by: AKIron on February 13, 2025, 04:34:50 PM
Of course you need to feed those LLMs. They get hungry.
Title: Re: DeepSeek
Post by: Brooke on February 14, 2025, 02:41:51 AM
Of course, Wistar's WISH was found to be HeLa contaminated like many others.
I think it's from biopsy from epithelial kidney cells from one living monkey in the 60's. Not from a bunch of monkeys who were killed and their kidneys minced. The cells from that monkey were from then on grown in labs and ended up mutating into what is used as Vero today (i.e., no longer regular monkey cells, but different in numerous major ways that make it suitable for other purposes and no longer the same as a monkey's cells).
In biology work, at a basic level, cells are just little machines that have different DNA and proteins in them. You alter the DNA and get cells that work for this or that purpose. Any process using cells -- not just some things with Vero, but making whiskey, beer, wine, yogurt, cosmetics, probiotic supplements, cheese, etc. -- is subject to getting contaminated with some organisms you don't want if you aren't careful.
Cell lines are part of the standard tool set in biology.
Title: Re: DeepSeek
Post by: icepac on February 21, 2025, 05:40:23 AM
The “vero of today” has benefitted from years of forcing mutations in vitro beyond the mutation that caused immortality to make it safe but it originally came from monkeys known to host SIV.
The dangers with using monkey kidney as a substrate were well known even in the 50s and some were flippant about the risks and weren’t worried about anything but beating polio to earn the nobel prize and secure funding forever.
Luckily, technology has come a long way toward “cleaning up dirty cell lines” but…….it still originated as minced monkey kidney.
Title: Re: DeepSeek
Post by: Brooke on February 21, 2025, 02:48:11 PM
The “vero of today” has benefitted from years of forcing mutations in vitro
Yep.
Quote
came from monkeys known to host SIV. ... The dangers with using monkey kidney as a substrate
Yes, SIV is a monkey virus that can infect monkey cells. What can also infect vero cells: measles, rubella, polio, flu, RSV, etc. It's why vero is used. Viral substrates have to be able to propagate viruses. It's their purpose. When you use such, obviously, you need to make sure it is infected with what you want and not with anything you don't want. This is a standard microbiology tool.
Like internal combustion engines use gasoline. But gasoline can light on fire, which can harm people. Yes -- but those features are also necessary for it to be a fuel. And you build engines and gas tanks such that they typically don't spray burning fuel all over. Does it still happen sometimes? Yes. But you use your standard engineering practices to make it rare enough such that gasoline and engines are viable to use.
Same goes for vero.
Quote
it still originated as minced monkey kidney.
Maybe. But I think it more likely the original source was a biopsy. Monkeys are expensive.