Latest / Strictly Business / The 5 o' Clock Shadow with Viv Govender

Transcript

The 5 o' Clock Shadow with Viv Govender · Strictly Business

• 0:00
  You're listening to Strictly Business Podcast with Lindsay Williams.
• 0:05
  It's Thursday, so it's time for the 5 o'clock shadow with Viv Govender and David
• 0:09
  Shapiro,
• 0:09
  except it's not the 5 o'clock shadow because it's not 5 o'clock yet and David
• 0:15
  Shapiro has cried off with some excuse.
• 0:17
  So Viv and I are going to tackle the world together, but who needs more than Viv
• 0:21
  Govender, especially Viv,
• 0:23
  when software companies have been under the spotlight so vigorously this week,
• 0:27
  haven't they?
• 0:27
  What on earth has happened?
• 0:28
  Well, I mean, I think...
• 0:29
  what's been happening is people are looking at what the future holds for AI,
• 0:35
  etc.,
• 0:35
  and are thinking that maybe we are seeing a situation in which some of these
• 0:40
  companies'
• 0:40
  future growth is not going to be what people think.
• 0:45
  I think, quite frankly, it may be over-predicting basically the impact of AI
• 0:51
  at the moment.
• 0:53
  But if you forecast into the future, I do think that it's reasonable to assume
• 0:59
  that in a couple of years' time, we are likely to see some impact of AI on these
• 1:05
  companies'
• 1:06
  revenue, etc.
• 1:06
  I saw an example.
• 1:08
  It was something to do with a legal software package,
• 1:12
  which analysed cases and illegal law over time periods.
• 1:16
  And they said that AI has completely replaced that.
• 1:19
  So this company was selling this software package, not out of business, but
• 1:25
  certainly having its business impacted.
• 1:27
  Is that something that's likely to be across the board?
• 1:28
  Yeah, I mean, for instance, I don't know why someone would use something like one
• 1:33
  of those language training programs, you know,
• 1:35
  when you could do it with the AI systems at the moment.
• 1:37
  And I think it gives you a better tutoring kind of result.
• 1:41
  I mean, some of the educational stuff I've already, I mean, I've like, you know, used
• 1:47
  Gemini and even
• 1:49
  ChatGPT recently to learn about new topics.
• 1:54
  You can go in depth in ways that you'd be You're pressed to find any kind of other
• 1:59
  system letting you get there.
• 2:01
  I mean, if you do physics, you can talk about quarks or Gemini at a level that
• 2:06
  you'd have to find a university professor to help you out with.
• 2:08
  But I mean, language stuff, you know, any kind of thing to do with, like you said,
• 2:12
  with the legal system, etc.
• 2:14
  These LLMs have gotten to a point right now where they are real competition for
• 2:18
  these programs.
• 2:20
  Talking about Gemini, and you've mentioned Gemini almost every time we speak, Viv.
• 2:23
  use Gemini, I use Copilot, I use Grok quite a lot.
• 2:26
  I also like DeepSeek.
• 2:27
  I don't know if you ever used DeepSeek, but did you see that article that I sent
• 2:33
  you from the Guardian newspaper in the UK,
• 2:35
  which looked at the personalities and the traits of each of the AI apps?
• 2:38
  It's very interesting to say one does this, one does that, one does the other.
• 2:42
  Did you agree with what you read, if you read it?
• 2:45
  I did, but I also think that, okay, look, I'm sure you've seen the French case
• 2:51
  against Grok.
• 2:53
  Yes.
• 2:54
  With regards to making lewd images, et cetera.
• 2:56
  You know, irrespective of the morality of it, I think there is a desire for this
• 3:02
  kind of stuff,
• 3:02
  or this stuff that goes against the terms and conditions of the providers out there.
• 3:07
  And I think what's going to happen in the future is that people are not going to be
• 3:13
  satisfied with accepting Grok's personality or Gemini's personality, et
• 3:15
  cetera,
• 3:17
  which is quite similar to what you talked about in the article.
• 3:19
  And what you're seeing with things like these malt pot.
• 3:20
  which we discussed, I think, a while back when it was called Cloudbot two weeks ago.
• 3:24
  And with Maltbot, what you'd like to find is people are going to basically have a
• 3:29
  personalized
• 3:31
  kind of thing that does what they want in the way they want it to be done.
• 3:34
  And if you can't find, for instance, like you said, a grok or an alphabet, et
• 3:41
  cetera,
• 3:41
  you'll use a deep secret, a Kimi, which is the new one from China,
• 3:45
  running your own server to do the things you want it to do in the ways you want it
• 3:49
  to do it.
• 3:49
  uh because uh if you asked grok or you asked uh you know chat gpt or whatever
• 3:56
  they would quite likely refuse
• 3:58
  you or maybe even call the authorities depending on what you're asking you to do
• 4:03
  you know what i mean um and it need not be something as as extreme as what
• 4:05
  grok's been accused of right now i mean uh there are For instance, certain
• 4:09
  restrictions on, you know,
• 4:11
  Holocaust denial, for instance, in France, that would obviously run afoul of the law
• 4:15
  there.
• 4:15
  But you could basically get a local system to do whatever you want with regards to
• 4:19
  that.
• 4:20
  So, yeah, I think that the future is not going to be you accepting Grok or you
• 4:26
  accepting Gemini's personality.
• 4:28
  It's you crafting a personality on your own local device, which may use Grok
• 4:33
  underlying it, or may use an open source system underlying it.
• 4:35
  I asked one of the apps, I said, can you please analyse the breakdown,
• 4:41
  can you give me a breakdown of the staff at Mar-a-Lago, how many, what their race
• 4:46
  is, what their gender is, etc.
• 4:48
  And it said, no, I'm sorry, we can't do that.
• 4:49
  And the reason I did it, of course, is because of Trump's recent pronouncements
• 4:54
  in the last few hours
• 4:56
  saying that a certain company has too many of this type of person and not enough of
• 5:01
  another type of person,
• 5:03
  which is blatant.
• 5:05
  stupid in my opinion but that's just me but this particular uh act wouldn't um
• 5:09
  wouldn't
• 5:11
  analyze it for me because of obviously the trump factor and also another thing i
• 5:16
  asked a simple question is
• 5:18
  there a market for this type of book covering this type of subject which
• 5:23
  because i was thinking of embarking on writing
• 5:25
  one and i got so many different answers from so many different sources one says
• 5:30
  yes fantastic idea this is how you do it the other one said no
• 5:32
  Yeah.
• 5:34
  No, nobody's going to read that.
• 5:36
  Very interesting, Viv.
• 5:38
  They've developed their own personalities.
• 5:40
  Yeah, most certainly.
• 5:42
  And look, also, they've learned what, remember, the way this works is that you
• 5:44
  get this thing trained to get this kind of lump of
• 5:46
  intelligence, and then that's kind of fine-tuned with human interaction.
• 5:47
  And over time, it learns what humans want it to say.
• 5:50
  So some things, like the earlier version of ChatGPT, were super kind of flattering,
• 5:56
  super kind of, you know,
• 5:58
  CQS, green with everything you said, so you say the world is flat as it, by God.
• 6:00
  You're such a genius to figure this out.
• 6:02
  The world needs to know more about your genius, et cetera, et cetera.
• 6:05
  And obviously the dangers that come from that.
• 6:07
  And I think that they still,
• 6:10
  that obviously the pleasing aspect that these things want to make sure they make
• 6:16
  you happy because that's what they've been trained on, not to be correct,
• 6:18
  but to make the person inputting questions happy or prompts happy.
• 6:23
  And so there is that danger.
• 6:24
  It does not give you the right answer when you do ask it for things.
• 6:28
  But like I said, I mean, I see the future and let me tell you not I mean not this
• 6:33
  year or next year but in a few years time
• 6:35
  I don't see why or how it's not a case where everybody has their own kind of um
• 6:40
  bolt port or whatever the version
• 6:42
  that is in the future an agent running that uses an underlying I know AI using
• 6:47
  multiple AIs underlying it but that
• 6:49
  bolt port is kind of tuned to what your level of sarcasm what your level of
• 6:51
  interest the
• 6:53
  way these things work is they need to have this huge context windows which has all
• 6:58
  the information it has about you uh to basically give you the right kind of on
• 7:00
  some.
• 7:02
  And I think that is something I think probably these different things don't
• 7:02
  have,
• 7:02
  because maybe Gemini has information about you in certain aspects that maybe ChatGPT
• 7:07
  doesn't.
• 7:09
  And that's why it's giving you different answers as to what it thinks will make you
• 7:10
  happy.
• 7:11
  Well, don't forget, of course, there was the very famous and very disturbing case
• 7:16
  of a South Korean woman
• 7:18
  who married her ChatGPT assistant.
• 7:20
  There was a ceremony.
• 7:21
  She wore a dress and everything else.
• 7:23
  And she was perfectly serious about this.
• 7:25
  she found that a human couldn't...
• 7:27
  be as nice and, as you said earlier on,
• 7:30
  obsequious as a chat GPT man or assistant did, whatever you call them.
• 7:37
  Very interesting.
• 7:38
  We have to get away from that.
• 7:40
  No, we don't get away from tech because the amount of volatility in this tech
• 7:46
  stocks rivals that of precious metals recently.
• 7:48
  I mean, I saw AMD doing one thing.
• 7:49
  I saw PayPal doing another the other day.
• 7:52
  And it's not uncommon to see, you know, 10, 15, 20 percent moves.
• 7:56
  in a day.
• 7:57
  I mean, that's pretty uncommon, I would say.
• 8:00
  Yeah, it is uncommon.
• 8:02
  But it's, again, a function of just, like you said, AMD, right, down, you know, 10
• 8:07
  plus percent.
• 8:09
  But if you look at the one year performance of this thing, it's up still
• 8:11
  78%.
• 8:13
  When you've done it before this, almost a 90% run in a year,
• 8:15
  the market basically needs to have some pretty strong numbers to justify keeping
• 8:19
  that up.
• 8:20
  And if you disappoint even slightly, a 10% pullback in a day is not impossible.
• 8:24
  I mean, another example I would put is Palantir.
• 8:25
  I mean, Yes, this share price is down dramatically in the last month or so, down
• 8:30
  20%.
• 8:31
  But if you give it a five-year chart on this thing, this thing is up three times
• 8:36
  plus in the last five years.
• 8:38
  Again, when you build in these kind of expectations, that's when you have the
• 8:40
  pullbacks.
• 8:40
  That's why we saw the same kind of thing happening with silver and gold.
• 8:42
  Yes, they came back dramatically, but the silver price fell after running up to
• 8:49
  about $120 an ounce.
• 8:51
  The gold price fell after going up to $5,500 an ounce.
• 8:53
  Silver was below $40 last year.
• 8:56
  gold was below 3,000 last year.
• 8:59
  When you do these kind of moves, these volatile pullbacks are what you'd expect
• 9:04
  irrespective of the underlying asset class.
• 9:06
  What do you make of precious metals as we get away from tech now?
• 9:08
  Because this time last week, when you, myself and David Shapiro were talking,
• 9:11
  the gold price actually went to a high $5,625 and that's within 24 hours, Viv.
• 9:18
  It was down below 4,500 briefly.
• 9:21
  It's bounced back to above 5,000, now below there again.
• 9:24
  It's all over the place.
• 9:25
  Silver did the same thing.
• 9:27
  And many people saying that the volatility in the speculative silver market led gold
• 9:33
  up and down and pulled it around a bit.
• 9:35
  But as you quite rightly say, when things do so well, of course, there's going to be
• 9:39
  corrections.
• 9:41
  But it does tell you that there is a big speculative bubble right at the top of the
• 9:46
  performance of the
• 9:48
  graphs of both AI stocks and precious metal stocks.
• 9:51
  Yeah.
• 9:52
  The only question is, are these things real?
• 9:53
  and the the the one thing I would say is look at what Google is doing at the moment
• 9:57
  with their CapEx.
• 9:59
  The market expected $119 billion of capex from Google.
• 10:01
  They're getting somewhere like 50% more than that, which is an insane number.
• 10:06
  The market basically obviously ran up on gold and silver, but why is it doing that?
• 10:10
  And I think the reason is fundamentally strong.
• 10:13
  I mean, there's two reasons to buy gold right at the moment.
• 10:15
  One is obviously the gold is a safe haven.
• 10:18
  The other is gold has demand for jewelry and other things.
• 10:22
  I think the safe haven reserve demand for gold is getting a stronger and stronger
• 10:27
  case going forward.
• 10:29
  Because of the fact that people are turning away from the US dollar, right now
• 10:31
  the US dollar is more than half of global reserves.
• 10:33
  No country I know should be happy with that.
• 10:35
  Even strong allies like Denmark, if you call it an ally anymore, or the EU, or
• 10:41
  Canada,
• 10:42
  but certainly not places like India or China.
• 10:45
  In those countries, I mean, you want to replace gold, the US dollar with something
• 10:51
  else, what are your choices?
• 10:53
  Is it the euro?
• 10:55
  That's almost the same as the US dollar because of the strong links those
• 10:55
  countries have.
• 10:57
  Gold and silver become the only things you can do.
• 10:58
  And unfortunately, there's not that much of that in the world.
• 11:00
  I mean, if you had to basically replace the US dollar entirely with gold and
• 11:06
  world's reserves,
• 11:08
  the gold price would be closer to 100,000 than to 5,000.
• 11:09
  You know what I mean?
• 11:11
  And the same thing applies to silver.
• 11:13
  So I think that's what's happening right now is that there is a fundamental shift.
• 11:17
  The future is more gold in the world's reserves than currently is there at the
• 11:23
  expense usually of the US dollar.
• 11:25
  That's going to result in a higher gold price.
• 11:25
  Whatever move happens in one day or happens over multiple years, I think is
• 11:29
  what the question is at the moment.
• 11:31
  The big question was over the last couple of years is will Bitcoin replace all these
• 11:37
  other instruments that you've
• 11:39
  just mentioned?
• 11:41
  Bitcoin plus other cryptocurrencies.
• 11:43
  There's a proliferation.
• 11:45
  I've got this website that I look at and it says cryptocurrencies.
• 11:47
  You click on it and it just keeps on going and going and going.
• 11:47
  These coins that you've never even heard of.
• 11:49
  Let's have a look at Bitcoin.
• 11:50
  Just creeping back up over 70,000 as we speak.
• 11:54
  It was 126,000 in October.
• 11:58
  And all the talk of Trump making the United States the crypto capital of the
• 12:05
  world drove it up
• 12:07
  a lot, I have to say, all that rhetoric.
• 12:07
  And Eric and Don Jr.
• 12:10
  going to Dubai and sitting on stages and sweating away and talking crypto up.
• 12:16
  It's all fizzled out.
• 12:17
  It's 45% lower.
• 12:19
  Is this the end of it or is this just like silver or gold or certain?
• 12:24
  tech stocks, it has to come down after such a run.
• 12:28
  Look, I mean, crypto is famous for these pullbacks.
• 12:30
  I mean, we saw, for instance, back in 2022 or something, it was down below 20,000,
• 12:37
  you know what I mean?
• 12:39
  And this was from being above 60,000 in 2021.
• 12:44
  And if you look at that kind of move, it's basically lost two thirds of the value.
• 12:47
  So if you look at going to 120, to do a similar kind of move there, you'd have to
• 12:53
  go below 40,000.
• 12:55
  to do a move of more than, say, two-thirds.
• 12:57
  Right now, I think my number is about 69,900 or so at the moment.
• 13:01
  Yes, exactly.
• 13:03
  And so I think, quite frankly, this is something to be expected with these
• 13:07
  currencies.
• 13:07
  I think they are, unfortunately at the moment, an entirely speculative asset
• 13:13
  class.
• 13:13
  There's no real use case for even Bitcoin, which is the most stable or the most
• 13:20
  established of these cryptocurrencies.
• 13:22
  If you're doing any transaction with crypto that you wouldn't want to do with
• 13:27
  the banking sector, my question is why are you doing that?
• 13:29
  And generally, it's because you're trying to evade some kind of laws or regulations.
• 13:32
  If you look at crypto's usefulness at the moment,
• 13:35
  the primary usefulness for transactions is either something that's semi or totally
• 13:41
  illegal, quite frankly.
• 13:43
  And that's sort of a great foundation going forward.
• 13:46
  And other than that, the main use is speculative.
• 13:48
  Basically, people buy it for the value to go up.
• 13:50
  It's not like it's like basically you don't have dollars in your account because
• 13:54
  you want the dollar value to go up.
• 13:56
  You have dollars in your account because you want to have, you know, some stable
• 13:59
  currency that you need to spend elsewhere.
• 14:01
  Until crypto becomes something like that there, unfortunately, these fluctuations
• 14:04
  are just something to be expected.
• 14:06
  Quickly, we must look at the U.S.
• 14:07
  jobs data which came out today.
• 14:08
  The ADP numbers showing a soggy, a soggy jobs market.
• 14:14
  Real jobs numbers, the non-farm payrolls, which we expect on the first Friday of
• 14:19
  every month,
• 14:19
  have been cancelled again because of shutdown problems in the United States of
• 14:24
  America.
• 14:24
  But also, I would imagine that the Bureau of Labor Statistics, its new people there,
• 14:28
  are massaging the numbers on behalf of the president.
• 14:31
  That's what a cynic would say.
• 14:33
  But anyway, U.S.
• 14:34
  jobs numbers, not particularly good.
• 14:36
  Before we go, Viv, any tech stories that we haven't covered, I introduced with the
• 14:42
  software woes.
• 14:44
  What have you seen in the last week since we spoke?
• 14:45
  Anything interesting?
• 14:46
  Well, speaking of those jobs numbers, I mean, I think half of the jobs cuts
• 14:53
  announced in January by like three companies was Dow, Inc.,
• 14:55
  UPS, and Amazon, which was basically the major things there.
• 14:59
  I mean, I think UPS was 30,000 jobs.
• 15:01
  Amazon was like 16,000 jobs.
• 15:03
  So I do think that's obviously and Dow, I think, was just under 5,000 jobs, so it
• 15:09
  wasn't as much.
• 15:10
  So these kind of things, I think, are quite...
• 15:12
  concentrated.
• 15:13
  There's questions as to whether or not this is AI related.
• 15:16
  I think with the Amazon one, there's a question not just whether it's AI
• 15:23
  replacing jobs or whether or not they need to cut the jobs to get free up enough
• 15:25
  money to invest in AI.
• 15:27
  And the other thing I think that we need to discuss, of course, or that's important
• 15:29
  to have noted, is those alphabet numbers.
• 15:31
  The CapEx spending in alphabet is just, I think, enormous by any means,
• 15:35
  approaching $200 billion over the next year.
• 15:38
  Countries don't do this, you know what I mean?
• 15:40
  This is not something even that China and than the US does on individual.
• 15:42
  Infrastructure prices, we're talking about building power systems or power grids or
• 15:47
  building highway systems.
• 15:49
  They don't build this kind of money.
• 15:51
  A private company is doing this.
• 15:53
  And the question I would say is this, if the guys with the most information and
• 15:56
  consider to be the smartest guy in the group,
• 15:58
  in the world basically, are accelerating the amount that they're spending on this
• 16:03
  technology, what does it mean?
• 16:05
  It's more likely not to mean that they are seeing things behind the scenes that we
• 16:08
  just don't see at the moment,
• 16:10
  probably showing that there's no wall, there's no plateau.
• 16:12
  And we are likely to see a very interesting next couple of years as the
• 16:17
  technology that they see right now.
• 16:19
  they give to us in a six months to a year's time or is it the fear of missing
• 16:24
  out because other companies are doing it they've got to do it and they've got to be
• 16:26
  seen to be
• 16:28
  the ones that are spending the most on capex is it is that a factor but
• 16:31
  potentially because at some stage people are going to say okay how
• 16:33
  much how much more sophisticated can artificial intelligence become without
• 16:37
  creating
• 16:39
  something that is better than the human brain which arguably they've already done
• 16:44
  yeah look i mean the thing with that
• 16:46
  i think it's interesting i don't know if i mention this to you or not but The
• 16:50
  multi-bot technology or the cloud-bot technology,
• 16:52
  it's the first time that you're seeing this wide-spect technology use that's
• 16:57
  actually making people use up their tokens and buy more tokens or
• 16:59
  upgrade their plans.
• 17:01
  I mean, if you just used AI and just asked a few questions, often you can get away
• 17:06
  with just using the free systems.
• 17:08
  And we can just go and basically use even the AI available on the Google search bar,
• 17:12
  you know what I mean?
• 17:14
  The AI mode there.
• 17:16
  But when you use these agents, when you use these things like Mallport, etc.,
• 17:21
  suddenly you find yourself using $100 of tokens in a day or maybe even more.
• 17:23
  And that scenario here, I think that is the killer app that gets you to the point
• 17:28
  in which people are using
• 17:30
  a lot more of this technology than they were in the past and spending more on it,
• 17:34
  more importantly, than they were in the past.
• 17:36
  And I think that if you were just trying to keep up with the Joneses, you might be
• 17:41
  willing to spend what was expected, $120 billion,
• 17:43
  maybe $130 billion, but not.
• 17:45
  190 billion?
• 17:47
  Why are you spending 50% more than the market expects?
• 17:48
  That's just not keeping up.
• 17:49
  And the next thing to note is that Google, on a different basis,
• 17:54
  is actually becoming a real player in the hardware space.
• 17:58
  Anthropic, as well as Apple, are likely to be using Google chips to help build their
• 18:03
  systems.
• 18:03
  And that means that if you look at the biggest hardware producer out there, which
• 18:08
  is obviously NVIDIA, or producer,
• 18:10
  but the player in the hardware space is NVIDIA, they don't actually produce, they
• 18:12
  just design.
• 18:13
  Does the company work more than Google?
• 18:15
  And I think that's also maybe indicating that maybe Google is seeing or Alford is
• 18:22
  seeing an opportunity here to encroach on a space that
• 18:24
  it's only been a bit player at the moment.
• 18:25
  It never stops, does it?
• 18:26
  There'll be something new next week.
• 18:27
  And next week, we will speak to you with David Shapiro.
• 18:30
  Viv, thank you very much for your time.
• 18:31
  That was Viv Govender from the award-winning RAN Swiss.
• 18:35
  And that was not the 5 o'clock shadow.
• 18:37
  The views and opinions expressed in these podcasts are those of Lindsay Williams and
• 18:42
  various contributors.
• 18:44
  and do not reflect the policy, position, or opinion of any other agency,
• 18:48
  organization, employer,
• 18:50
  or company associated with StrictlyBusinessPodcast.com.
• 18:53
  Assumptions made on the analyses are not reflective of the position of any other
• 19:00
  entity other than the speaker or the author.
• 19:02
  And since we are critically thinking human beings, these views are always subject to
• 19:06
  change, revision,
• 19:08
  and rethinking at any time.
• 19:08
  Please do not hold us to them in perpetuity.