Latest / Connecting the Dots / Connecting the Dots - 6-4-2025

Transcript

Connecting the Dots - 6-4-2025 · Connecting the Dots

• 0:00
  Today's tech news moves fast, but understanding
• 0:03
  it shouldn't be hard. That's why Snarful Solutions
• 0:06
  Group brings you Connecting the Dots, where we
• 0:09
  break down what's happening in tech and why it
• 0:11
  matters to you. Now, let's dive in with your
• 0:14
  hosts. Welcome back to the Deep Dive. Glad to
• 0:17
  be here. Our mission here, it's simple really.
• 0:20
  You hand us your stack of sources. Could be articles,
• 0:23
  research notes, anything. And we dive right in.
• 0:26
  We dig out. the most important knowledge, those
• 0:30
  surprising facts, the crucial insights. And put
• 0:32
  it all together so you can get well -informed,
• 0:34
  fast, you know, without feeling buried by it
• 0:37
  all. Think of this as guides, really. Sifting
• 0:39
  through the material you give us, helping you
• 0:41
  see what's essential, understand the context.
• 0:43
  And connect the dots, make it stick. Exactly.
• 0:46
  So today, we've got a really interesting mix,
• 0:48
  sources covering three areas. Seemingly distinct,
• 0:52
  but definitely connected. Finance and tech. Yep.
• 0:56
  We're going to unpack reports on, first, the
• 0:59
  early performance of a new stable coin, a high
• 1:02
  -profile one. Then the crazy fast competition
• 1:05
  among the big AI models. It's intense. And finally,
• 1:08
  some pretty significant government investments,
• 1:11
  trying to boost computer chip manufacturing right
• 1:13
  here in the US. Right. So enough talk about the
• 1:16
  deep dive. Let's just dive in. We'll jump into
• 1:18
  these sources together, figure out what's really
• 1:19
  going on, and crucially, why it matters to you.
• 1:22
  Let's unpack this. Agreed. Let's start with that
• 1:25
  first topic from your sources. USD1, the launch,
• 1:29
  the early days. It's from President Trump's World
• 1:32
  Liberty Financial. Okay, USD1. Now, for anyone
• 1:36
  maybe not deep in this space, let's quickly level
• 1:38
  set. The sources do explain what a stablecoin
• 1:40
  is. Yeah, basically it's a digital currency,
• 1:42
  right? Designed to hold a stable value, usually
• 1:45
  pegged to something like the US dollar. So it
• 1:47
  acts like a tool for crypto trading. Exactly.
• 1:50
  A key tool. Let's traders move between different
• 1:52
  cryptos without always converting back to like
• 1:55
  regular money. Yeah, digital dollars, essentially.
• 1:57
  And this isn't some tiny niche. The sources point
• 2:00
  out the scale, the whole stable coin market over
• 2:02
  two hundred and thirty seven billion dollars.
• 2:04
  Oh, absolutely massive. You've got giants like
• 2:06
  Tether at one hundred and fifty one billion dollars
• 2:08
  circle around sixty billion dollars. So launching
• 2:10
  USD one. Well, you're up against some heavy hitters.
• 2:14
  Definitely. And the initial data. According to
• 2:17
  these sources. It suggests USD one had kind of
• 2:20
  a sluggish start. That's what they're saying
• 2:22
  muted inflows Not a lot of organic demand right
• 2:25
  out of the gate. But what's interesting is where
• 2:28
  it did see some action early on the source has
• 2:31
  mentioned decentralized exchanges, right like
• 2:33
  PancakeSwap. Yeah, PancakeSwap, which runs on
• 2:36
  Binance's Smart Chain. Apparently, after USD1
• 2:38
  got listed on Binance back in May, the daily
• 2:41
  on -chain volumes on PancakeSwap jumped, topped
• 2:44
  $14 million, according to Kyco data, in one source.
• 2:47
  Okay, $14 million there. But the contrast is
• 2:50
  key, isn't it? Volume on Binance itself, the
• 2:52
  big centralized exchange, it was lower. Much
• 2:55
  lower. Lagged behind at around $8 million. And
• 2:59
  this leads to that analysis from Kyco that the
• 3:01
  source has mentioned. Right. Their analysts basically
• 3:03
  said USD1 hasn't really made the leap. It hasn't
• 3:06
  jumped from those decentralized platforms, which,
• 3:08
  let's face it, cater to a more niche crypto crowd
• 3:11
  to the big centralized exchanges serving the
• 3:14
  wider market. They called it success in a niche
• 3:16
  market of a niche market. Kind of sums it up.
• 3:19
  A niche of a niche, okay. And what about the
• 3:20
  liquidity? Is it genuine users? Well, that's
• 3:23
  the question. The Kyco data site shows over half
• 3:26
  the USD1 liquidity on PancakeSwap comes from
• 3:29
  just three wallets. Three wallets, wow. Yeah,
• 3:31
  and analysts suspect these are probably market
• 3:33
  -making wallets, likely tied to World Liberty
• 3:36
  Financial itself. You know, providing liquidity
• 3:39
  to make trading possible rather than showing
• 3:41
  lots of people want to hold it. So maybe not
• 3:43
  organic demand just yet. More like supporting
• 3:46
  its own market. Potentially, yeah. And the sources
• 3:48
  explore why it might be lagging behind rivals
• 3:51
  like Tether, Circle, even Ripple's new RLUSD,
• 3:55
  which apparently hit $50 million daily volume
• 3:58
  on centralized exchanges. And the reasons cited.
• 4:00
  Seems to be a lack of big institutional partners,
• 4:03
  maybe. Or those kinds of promotional pushes you
• 4:05
  often see to get early traction. The sources
• 4:08
  suggest USD1 didn't really lean into those strategies,
• 4:11
  at least initially. Interesting. Now, stepping
• 4:14
  back a bit. One source quotes Donald Trump Jr.
• 4:17
  giving a sort of strategic view. Right. He frames
• 4:20
  stablecoins as more than just business, as a
• 4:22
  way to reinforce the US dollar's dominance globally,
• 4:26
  called them the savior of dollar hegemony. Strong
• 4:28
  words. Very strong. And the sources also note,
• 4:32
  impartially, World Liberty Financial's other
• 4:35
  activities, like a native token, WLFI. Yeah,
• 4:39
  WLFI. They reportedly raised over $550 million
• 4:43
  dollars selling that token. Sources mentioned
• 4:46
  75 % of the profits go to family entities, and
• 4:49
  the whole venture started just before Trump's
• 4:51
  election win. And there's mention of broader
• 4:53
  scrutiny, too, around Trump's crypto stuff, like
• 4:56
  the Trumpy meme coin. The sources touch on that.
• 4:59
  Controversies like a VIP contest for top holders.
• 5:02
  Senator Warren's criticism calling it an orgy
• 5:04
  of corruption. Again, just reporting what the
• 5:06
  sources say. And data showing big gains for a
• 5:09
  few wallets, while many others lost money collectively.
• 5:11
  It's all in there, presented neutrally. Exactly.
• 5:13
  Now, back to USD1 itself. The structure. How's
• 5:16
  it backed? Sources say it's fully backed. U .S.
• 5:19
  Treasuries, dollars, cash equivalents, held by
• 5:21
  Big Go Trust. OK. And World Liberty says it'll
• 5:24
  be audited regularly by a third party. Though
• 5:26
  the sources didn't have the specific details
• 5:28
  on which firm or how often. Their stated goal
• 5:30
  is pretty ambitious, though. Allow sovereign
• 5:33
  investors and major institutions to use it for
• 5:37
  seamless, secure, cross -border transactions.
• 5:40
  High aims. But an industry expert quoted Kevin
• 5:42
  Latineiti, kind of brings it back to earth. He
• 5:45
  says launching one is easy. The classic launch
• 5:48
  is easy. Adoption is hard. Precisely. Building
• 5:51
  the ecosystem, getting actual usage. That's the
• 5:53
  real challenge. He points out the stiff competition
• 5:56
  and asks, you know, is this competing with other
• 5:58
  U .S. businesses or partnering? An interesting
• 6:00
  question. It is. Now, despite that slow organic
• 6:04
  start, there was a recent tech development mentioned.
• 6:06
  Integration with Chainlink. CCIP. Yes. Announce
• 6:10
  at Consensus 2025 Chainlink's cross -chain interoperability
• 6:14
  protocol. CC IP and that's significant because
• 6:17
  it let the USD one work across different block
• 6:20
  chains securely starting with the theorem and
• 6:22
  BNB chain more planned This tackles a huge security
• 6:25
  headache in crypto those cross -chain bridges
• 6:27
  have been notoriously vulnerable right sources
• 6:30
  mentioned nearly three billion dollars lost historically
• 6:33
  through bridge hacks Exactly. So World Liberty's
• 6:36
  co -founders are quoted saying Chainlink offers
• 6:38
  institutional -grade security and is absolutely
• 6:41
  critical for merging TradFi and DeFi. It's a
• 6:44
  tech step forward. Okay, and there was that big
• 6:46
  transaction. $2 billion involving an Abu Dhabi
• 6:49
  fund? Yeah, MGX, yeah. The sources say that investment
• 6:53
  in Binance used USD1 and it helped USD1 hit a
• 6:56
  $2 billion market cap. But the Kaiko analyst,
• 7:00
  McCarthy, noted that even that huge deal didn't
• 7:03
  really boost USD1's visibility or on -chain activity
• 7:06
  much. The velocity didn't increase as you might
• 7:09
  expect. So even a massive transaction didn't
• 7:12
  spark broader use. Seems that way. So the takeaway
• 7:15
  for you, the learner, why track this stable coin's
• 7:17
  journey? Well, I think it really shows those
• 7:20
  market adoption hurdles, right? Even with big
• 7:22
  names, getting real users is tough. Absolutely.
• 7:24
  It's a case study in launching digital assets.
• 7:27
  The tech matters, like that CCIP integration
• 7:29
  for bridging TradFi and DeFi. But tech alone
• 7:33
  isn't enough. You need the ecosystem, you need
• 7:35
  organic demand. It's a clear example of technology
• 7:38
  moving fast, but market traction being a whole
• 7:40
  different challenge. Okay, let's switch gears
• 7:43
  now. Let's pivot to that second area from the
• 7:45
  sources. The AI model race. It's moving incredibly
• 7:50
  fast, way beyond just chat GPT now. Oh yeah,
• 7:52
  chat GPT is still the leader for sure. The source
• 7:55
  has confirmed that over half a billion weekly
• 7:57
  active users back in late March. Still huge.
• 8:00
  But get this, the rivals are definitely gaining
• 8:02
  ground. The data from similar web and sensor
• 8:05
  tower from March Feb this year is pretty revealing.
• 8:08
  Like Google Gemini. It hit 10 .9 million average
• 8:11
  daily web visits in March. That's up 7 .4 % just
• 8:14
  from February. And the Gemini app saw a big jump
• 8:17
  too after that 2 .0 flash release. A 42 % increase
• 8:20
  in weekly active users. That's significant growth.
• 8:23
  And anthropics clawed. They're not sleeping either.
• 8:25
  Nope. 3 .3 million average daily web visits in
• 8:28
  March, and its app users were up 21 % week over
• 8:31
  week after they dropped the 3 .7 sonnet version.
• 8:34
  Microsoft Copilot, using OpenAI's tech, also
• 8:36
  saw growth, though maybe a bit steadier. Yeah,
• 8:38
  2 .4 million daily web visits, up about 2 % from
• 8:41
  Feb. But then you have some maybe less expected
• 8:44
  players making noise. Like DeepSeek, Chinese
• 8:47
  lab. Right. Sources show it hit 16 .5 million
• 8:51
  average daily web visits in March. Although,
• 8:54
  interestingly, that was actually down 25 % from
• 8:57
  February, so volatile maybe? And the big surprise
• 9:00
  mover, XAI's Grok. Huge momentum there. Also
• 9:04
  averaged 16 .5 million daily visits in March,
• 9:07
  but the growth was insane, nearly 800 % month
• 9:10
  over month. Feb to March. Wow, 800%. Okay, so
• 9:14
  lots of action. But the sources add important
• 9:16
  context about mobile, right? Crucial context.
• 9:19
  As of March, ChatGPT still had 10 times the mobile
• 9:22
  app weekly active users compared to Gemini and
• 9:24
  Claude put together. 10 times. So the web traffic
• 9:27
  might be shifting, but mobile dominance is still
• 9:29
  firmly with ChatGPT. For now, yeah. And experts
• 9:32
  cited say this growth isn't just about new models
• 9:34
  dropping. It's also new features, better capabilities.
• 9:37
  Like Gemini adding coding tools, Claude integrating
• 9:39
  other functions. Exactly. Plus just growing consumer
• 9:41
  interest. People are finding more unique ways
• 9:43
  to actually use these. things now. And speaking
• 9:45
  of new models, Meta made a big splash recently.
• 9:48
  The sources cover Llama 4. Yeah, their latest
• 9:51
  models now integrated into Meta AI assistance
• 9:53
  on WhatsApp, Messenger, Instagram, the web, big
• 9:57
  distribution. And there's a whole family. Scout,
• 10:00
  Maverick. And Behemoth. Scout's the lightweight
• 10:03
  one. Maverick is bigger, tougher. Behemoth is
• 10:05
  the top tier, but it's still training, not out
• 10:07
  yet. Scout and Maverick are available now. And
• 10:09
  Meta is making some bold claims based on internal
• 10:11
  tests, according to the source. Valor. Claiming
• 10:14
  llama for Scout beats Google Gemma 3 smaller
• 10:17
  Gemini versions Mistral 3 .1 and Maverick They
• 10:21
  say it's on par with or even better than GPT
• 10:24
  4 .0 and Gemini 2 .0 flash in some areas like
• 10:27
  coding reasoning That's the claim. Coding, reasoning,
• 10:31
  multilingual, long -context, image generation.
• 10:33
  Recommended for creative writing, too. However...
• 10:36
  There's always a however. The sources note Metta's
• 10:38
  own test showed it struggled against the really
• 10:40
  high -end models. Gemini 2 .5 Pro, GPT 4 .5,
• 10:44
  Claude 3 .7 Sonnet, still a gap there. Got it.
• 10:47
  And Behemoth, the one still training. They're
• 10:49
  calling it potentially the highest performing
• 10:51
  base model in the world. Expecting it to be...
• 10:54
  GPT 4 .5 and GLOD Sonnet 3 .7 on some technical
• 10:58
  stuff, but maybe not Gemini 2 .5 Pro. Sources
• 11:01
  also flag the high hardware cost for training
• 11:04
  it. All these use that mixture of experts, MoE,
• 11:07
  architecture. Right. Designed for efficiency.
• 11:10
  Only activating the expert parts needed for a
• 11:12
  task. Now that open source label meta uses. The
• 11:17
  sources clarify that, don't they? They do. It
• 11:18
  has licensing limits. Big companies over 700
• 11:21
  million active users need special permission.
• 11:25
  So by the strict open source initiative definition
• 11:27
  mentioned, it's not truly open source. An important
• 11:30
  distinction. Yeah. And competitors are influencing
• 11:32
  each other too. Meta looking at DeepSeek. Apparently,
• 11:35
  yeah. DeepSeek got attention for being competitive
• 11:37
  but cheap to run. Sources say Meta's studying
• 11:40
  their methods, possibly looking to cut their
• 11:42
  own operating costs. Competition drives innovation,
• 11:45
  right? And OpenAI. Quick note on GPT -5. Timeline
• 11:48
  adjusted slightly, expected in coming months
• 11:49
  now, with some smaller reasoning models, 03 and
• 11:52
  a 4 Mini, coming sooner. Okay, so this crazy
• 11:55
  competitive landscape, all these releases, what's
• 11:58
  the takeaway for you, the listener? I think the
• 12:00
  main thing is just the speed. This intense race
• 12:03
  means faster innovation, better features, more
• 12:06
  specialized tools coming online constantly. And
• 12:09
  companies are competing differently, not just
• 12:10
  raw power, but efficiency, integration, like
• 12:13
  meta strategy. Exactly. It means you really need
• 12:16
  to stay informed about what each model does best,
• 12:19
  where it's available, because the capabilities
• 12:21
  are starting to diverge more. It's not one size
• 12:23
  fits all anymore. A fast moving target keeps
• 12:26
  us on our toes. Definitely. All right. Final
• 12:28
  shift. Let's move to our third topic from the
• 12:31
  sources. Boosting US chip production. Big government
• 12:34
  money involved here. Huge news highlighted in
• 12:37
  the sources, yeah. The Biden administration giving
• 12:39
  $1 .5 billion to Global Foundries, GF. And this
• 12:43
  is part of that CHPS and Science Act from 2022.
• 12:46
  That's the one. This is the third direct funding
• 12:48
  award under that act. The goal is over $52 billion
• 12:51
  overall to revitalize U .S. chipmaking and R
• 12:54
  &D. And $1 .5 billion direct funding. Plus loans
• 12:59
  available. Right. Up to $1 .6 billion in potential
• 13:02
  loans on top of that. So combined with what GF
• 13:06
  invests, the sources say the total public -private
• 13:08
  investment here is looking like roughly $12 .5
• 13:12
  billion. Wow. That's serious money. Where's it
• 13:15
  going? What are they building? Three main things
• 13:17
  outlined in GF's plans. a brand new advanced
• 13:20
  chip factory in Malta, New York, expanding production
• 13:23
  at their existing plant in Malta, that's part
• 13:25
  of a deal they have with GM, and revitalizing
• 13:29
  their plant in Burlington, Vermont. And the importance
• 13:31
  of these chips, Commerce Secretary Raimondo called
• 13:33
  them essential. Absolutely essential. For military
• 13:36
  gear, EVs, smartphones, faster internet, critical
• 13:41
  for national security, she stressed. Which ties
• 13:43
  into those broader concerns mentioned. Senator
• 13:45
  Schumer was quoted saying chips are basically
• 13:47
  like food for the economy and national security.
• 13:49
  Yeah, pretty stark comparison. He tweeted about
• 13:51
  the GF award being NY's first, the nation's biggest
• 13:55
  so far under the act, directly linked it to jobs,
• 13:58
  national security, helping the auto industry.
• 14:00
  And the sources remind us why this is happening
• 14:02
  now. The pandemic really exposed the vulnerability,
• 14:05
  didn't it? Painfully so. That reliance on foreign
• 14:08
  chips caused shortages everywhere, auto plants
• 14:11
  especially. Experts estimated it cost the U .S.
• 14:14
  economy $240 billion, highlighted a massive weakness.
• 14:19
  And if we still don't make that many here, what
• 14:20
  are the numbers? Currently, sources say the U
• 14:22
  .S. produces only about 12 % of the world's chips.
• 14:26
  Heavily reliant on Asia, TSMC in Taiwan makes
• 14:29
  90 % of the world's most advanced chips. 90%.
• 14:32
  That's staggering. Now... The sources touch on
• 14:35
  the politics, too, which we'll just report. Democrats
• 14:37
  emphasize preventing rivals like China, Russia
• 14:40
  from gaining advantage, long -term investment.
• 14:43
  Right. Framing it is vital for the future, showing
• 14:45
  Americans we have a future, and pointing to the
• 14:47
  bipartisan support the CHPS Act had initially.
• 14:50
  While Republicans, in the election context, focus
• 14:52
  more on immediate economic concerns like inflation
• 14:55
  impacting buying power. different focuses presented
• 14:57
  in the sources. Exactly. Now, the jobs impact.
• 15:00
  What do the sources estimate? They anticipate
• 15:02
  1 ,500 manufacturing jobs from these GF projects,
• 15:05
  plus 9 ,000 construction jobs over the next decade.
• 15:08
  Significant numbers and some worker benefits
• 15:10
  mentioned too. Yeah, $10 million for worker training.
• 15:13
  And GF is extending its $1 ,000 annual child
• 15:18
  care subsidy to the construction workers on these
• 15:20
  projects, which is interesting. Good detail.
• 15:22
  And the market liked the news. Gee, of stock.
• 15:25
  Sources say it was up noticeably after the grant
• 15:27
  announcement. So investors saw it positively.
• 15:31
  And they mentioned previous smaller GPS awards
• 15:33
  to BAE systems, microchip technology. Right.
• 15:37
  So this massive investment in bringing chip making
• 15:40
  home. Why does this matter to you, listening
• 15:43
  right now? Well, it seems pretty fundamental.
• 15:45
  It's about making sure the stuff we use every
• 15:47
  day, phones, cars, computers, isn't so vulnerable
• 15:50
  to global disruptions, right? Supply chain resilience.
• 15:53
  That's a huge piece of it. Plus, job creation
• 15:56
  here in the U .S. and national security, as mentioned.
• 15:58
  And it's really about laying the groundwork for
• 16:00
  future tech innovation. You need domestic chip
• 16:03
  capacity for that. Absolutely. Foundational stuff,
• 16:05
  from security to future tech. Okay, so let's
• 16:08
  just quickly recap the three big areas we covered
• 16:10
  today. based on those sources you gave us. First,
• 16:12
  we looked at USD1, the new stablecoin. Its early
• 16:16
  struggles, the tech moves like CCIP integration,
• 16:19
  and really the challenge of market adoption,
• 16:22
  even with a big name attached. Then we dove into
• 16:26
  that AI arms race. ChatGPT's still leading, but
• 16:29
  Gemini, Claude, Grok, Metis Llama 4, everyone's
• 16:33
  pushing hard. Faster features, different strategies
• 16:35
  emerging. It's intense. And finally, that big
• 16:39
  push for US chip manufacturing, the Global Foundry's
• 16:42
  investment under the CH ICE Act, driven by supply
• 16:45
  chain worries, national security, and future
• 16:47
  tech needs. Understanding these threads, even
• 16:50
  from this focused dive, hopefully gives you a
• 16:52
  clearer picture of these key trends shaping tech
• 16:54
  and finance, and how they're all kind of tangled
• 16:56
  together. Definitely not isolated issues, they
• 16:58
  all influence each other. Which leads us to a
• 17:00
  final thought, something for you to mull over.
• 17:02
  As finance gets more digital, AI gets more powerful,
• 17:05
  and global supply chains keep getting reshaped.
• 17:07
  How will these dynamics we discussed, the push
• 17:09
  for digital currency adoption, the fight for
• 17:11
  AI supremacy, the drive for domestic tech manufacturing,
• 17:15
  how will all that together reshape the global
• 17:17
  economic and geopolitical map in the coming years?
• 17:20
  That's a big question. These sources definitely
• 17:22
  give us clues about the forces at play. We really
• 17:25
  hope this deep dive helped you cut through some
• 17:27
  noise, find those key insights. Absolutely. Thanks
• 17:31
  for joining us. If any of this sparked more questions,
• 17:34
  or hey, if you have sources on a different topic
• 17:36
  you want us to tackle, let us know. We built
• 17:39
  this show for you. We love helping you dig through
• 17:41
  the information overload. Until the next deep
• 17:43
  dive, keep digging. Thanks for listening and
• 18:01
  we'll catch you next time.