Latest / Connecting the Dots / Connecting the Dots - 5-27-2025

Transcript

Connecting the Dots - 5-27-2025 · 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 hosts.
• 0:16
  Hey everyone, hope you had a good restful Memorial
• 0:18
  Day weekend. Hope you did too. It's Tuesday,
• 0:21
  May 27th, 2025, and you're listening to Connecting
• 0:25
  the Dots. We're coming to you from Snarful Solutions
• 0:28
  Group. right here in Sacramento, California.
• 0:30
  Yep. And it is a sunny one today, looking like
• 0:33
  a high of about 90 degrees. Nice day for it.
• 0:36
  Definitely. Before we jump into the tech news,
• 0:40
  just a quick look back at the markets from last
• 0:42
  Friday. That was May 23rd. The Dow Jones closed
• 0:45
  at 41603 .07 and the S &P 500 finished at 5802
• 0:51
  .82. Okay. So with that, let's get into our deep
• 0:55
  dive. We've got three big tech stories today
• 0:57
  and we'll connect the dots on how they might
• 0:59
  impact things. Sounds good. First up, pretty
• 1:02
  significant acquisition making waves. Salesforce.
• 1:05
  Ah, Salesforce. What are they up to? They've
• 1:07
  announced they're buying own company. It's a
• 1:09
  $1 .9 billion deal, all cash. $1 .9 billion.
• 1:13
  Okay. Own company. They're in data management.
• 1:15
  Right. Cloud data. Exactly. Cloud data management
• 1:17
  and protection. They're based out of New Jersey.
• 1:19
  And this is what? Salesforce's biggest buy since
• 1:21
  Slack? Yeah, biggest in Slack back in 2021. It's
• 1:25
  kind of an interesting timing, too. How so? Well,
• 1:27
  you remember earlier this year, there was all
• 1:29
  that talk about them maybe acquiring Informatica.
• 1:32
  Oh, right. That was a much bigger potential deal,
• 1:34
  but it fell through. It did. So this acquisition
• 1:37
  of OWN feels maybe more focused, more strategic,
• 1:41
  perhaps. Makes sense. What's Salesforce saying
• 1:43
  about it? Well, Salesforce GM Steve Fisher put
• 1:46
  out a statement really emphasizing their commitment
• 1:48
  to secure end -to -end data protection. So doubling
• 1:52
  down on security in the cloud. Seems critical
• 1:55
  these days. Absolutely. And they actually started
• 1:57
  back in 2015 as Own Backup. They provide tools,
• 2:02
  really vital tools, for things like automated
• 2:04
  backup, disaster recovery, mainly for SaaS applications.
• 2:08
  Like Salesforce itself, I assume? Yep. Salesforce,
• 2:11
  but also AWS, Microsoft platforms, the big cloud
• 2:15
  players. They had raised a good amount of funding,
• 2:17
  over $500 million. Wow. And they were valued
• 2:19
  pretty highly a couple years back, around $3
• 2:21
  .35 billion. There's clearly a big market need
• 2:26
  here. The whole data backup and recovery sector
• 2:29
  seems to be booming. Oh, definitely. It was worth
• 2:31
  what? Almost $13 billion in 2023? Yeah, nearly
• 2:34
  $13 billion and growing fast to like an 11 %
• 2:37
  average growth rate year over year for a while
• 2:40
  now. And you can see why, right? With ransomware
• 2:43
  threats constantly in the news, data center disasters.
• 2:45
  Remember that big OVH fire in France? Yeah, that
• 2:48
  was major. And then you've got increasing regulations
• 2:51
  like the EU AI Act. putting more pressure on
• 2:54
  how companies handle data. Right, so own company
• 2:56
  fits right into that picture. How many customers
• 2:58
  do they have? They're saying nearly 7 ,000 customers.
• 3:01
  And it's not just backup anymore. They do data
• 3:04
  archiving, seeding, security, even analytics
• 3:07
  now. Pretty comprehensive. Yeah. and own CEO
• 3:09
  Sam Gutman basically said their tools really
• 3:12
  complement what Salesforce already offers, makes
• 3:15
  the whole data protection picture stronger for
• 3:17
  users. So when does this deal actually close?
• 3:20
  They're expecting it to close in Q4 of Salesforce's
• 3:23
  fiscal year 2025. OK, so you mentioned the Informatica
• 3:27
  thing falling through. Does this one point nine
• 3:30
  billion dollar deal maybe signal Salesforce is
• 3:33
  getting back into larger M &A even after they
• 3:35
  sort of suggested they were backing off big deals?
• 3:37
  It could. It might mean they're just being more
• 3:40
  selective targeting key areas like data protection.
• 3:43
  They did make a couple of smaller acquisitions
• 3:45
  recently to predict spring and 10x. Right. So
• 3:48
  maybe a more focused strategy now. Possibly.
• 3:51
  Yeah. OK. Moving from the cloud. Let's go way,
• 3:54
  way up to space. SpaceX seems to be making noise
• 3:58
  about Mars again. Yeah, it sounds like they're
• 4:01
  really accelerating efforts for a potential Mars
• 4:03
  mission. Uncrewed, mind you. Uncrewed, right.
• 4:05
  And possibly as early as 2026. That seems to
• 4:08
  be the target they're pushing for. Reports suggest
• 4:10
  they're even moving people around internally.
• 4:12
  Like shifting resources. Exactly. Moving employees
• 4:15
  from the Dragon program, which handles crew and
• 4:18
  cargo to the space station, over to Starship
• 4:20
  development to speed things up. That's a big
• 4:22
  commitment, especially after those, well, less
• 4:24
  than perfect Starship test flights earlier this
• 4:27
  year. Yeah, the ones in January and March 2025
• 4:29
  didn't go quite as planned. They were pretty
• 4:33
  high profile failures or maybe rapid unscheduled
• 4:36
  disassemblies, as they sometimes say. Aha, right.
• 4:39
  But it clearly hasn't deterred them. Didn't the
• 4:41
  FAA just give them the green light for more launches?
• 4:44
  They did. A big increase, actually. Up from five
• 4:46
  Starship launches a year to 25. 25. Wow. And
• 4:50
  they expanded the hazard zones around the launch
• 4:52
  site, too. One of their engineering execs, Shana
• 4:54
  Dees, acknowledged on X it was a rough start
• 4:56
  of the year for Starship. But, yeah, they're
• 4:58
  pushing ahead. And there's a military angle,
• 5:00
  too, isn't there? Starfall. Right, Starfall.
• 5:03
  That's their project for potentially moving large
• 5:05
  amounts of government cargo, like over 66 ,000
• 5:07
  pounds. They're expecting more funding for that,
• 5:10
  about 149 million dollars. So government interest
• 5:12
  is definitely there in this 2026 Mars timeline.
• 5:18
  Why then, specifically? Well, partly it's about
• 5:21
  orbital mechanics. Earth and Mars align favorably
• 5:24
  for launches about every 26 months. 2026 is the
• 5:27
  next window. Okay, the launch window. And it
• 5:29
  also happens to align with some proposed NASA
• 5:32
  budget shifts under the Trump administration,
• 5:35
  potentially an extra billion for Mars efforts,
• 5:37
  though other NASA programs might face cuts. Interesting
• 5:40
  timing, but technologically it's still a huge
• 5:42
  leap, right? What are the main hurdles? Oh, massive
• 5:46
  hurdles remain. In -orbit refueling is a big
• 5:48
  one. They think they'll need three separate refueling
• 5:51
  missions in orbit just to get Starship to Mars.
• 5:54
  Three refuels in space, that's complex. Very.
• 5:57
  Plus the lunar version of Starship... The one
• 5:59
  NASA needs for its Artemis mission to the moon
• 6:01
  in 2027, it's already facing delays. Right. So
• 6:05
  a lot needs to go right. But this is core to
• 6:07
  Musk's vision, isn't it? Mars colonization. Absolutely.
• 6:09
  That's why he started SpaceX. He even mentioned
• 6:12
  maybe putting Tesla's Optimus robot on that 2026
• 6:14
  flight. A robot on the first flight. That's the
• 6:17
  idea. And he's aiming high, talking about sending
• 6:19
  five uncrewed starships to Mars by 2026. Five.
• 6:23
  Five. And if those go well, then maybe crewed
• 6:26
  missions during the next window, which would
• 6:28
  be late 2028 or early 2029. That feels incredibly
• 6:32
  ambitious, especially since, as you mentioned,
• 6:35
  they haven't actually landed Starship back on
• 6:37
  solid ground yet. Correct. The last test flight
• 6:40
  ended with a splashdown in the Indian Ocean,
• 6:42
  which was progress, but Nautiland had landing.
• 6:46
  Still, Musk talks about exponentially increasing
• 6:48
  Mars trips, aiming for a self -sustaining city
• 6:51
  in, what, 20 years? Yeah, that's the long -term
• 6:54
  vision he paints. Though it's fair to say his
• 6:56
  timelines for Mars have shifted over the years.
• 6:59
  Understood. And his voice concerns about bureaucracy
• 7:02
  slowing things down. He has. Worries about regulations
• 7:05
  hindering the Starship program. We've already
• 7:08
  seen NASA's Artemis 3 mission, which needs Starship,
• 7:11
  get delayed to 2026. And that private lunar trip
• 7:14
  by the Japanese billionaire Maezawa got canceled
• 7:17
  too. Right. Just quickly, for anyone maybe less
• 7:19
  familiar, can you remind us what Starship actually
• 7:21
  is? Sure. It's basically two parts. The Super
• 7:24
  Heavy Booster, which is the first stage, and
• 7:26
  then the Starship spacecraft itself, the upper
• 7:28
  stage. It's about 165 feet tall. Together, it's
• 7:31
  the biggest, most powerful rocket ever built.
• 7:34
  Biggest ever. And the key design goal is full,
• 7:37
  rapid reusability. They plan to eventually land
• 7:40
  the super -heavy booster right back on the launch
• 7:43
  tower using those big arms they call chopsticks.
• 7:45
  Catching the booster with arms? Wild! How many
• 7:49
  tests have they done? Four main flight tests
• 7:51
  so far, each one getting further and achieving
• 7:53
  more objectives. The next one, flight five, is
• 7:57
  supposed to be the first attempt at catching
• 7:58
  the booster with those chopsticks. Okay, definitely
• 8:01
  want to watch. Let's shift gears for our last
• 8:03
  story. From hardware in space to software and
• 8:07
  talent here on Earth, AI talent specifically.
• 8:10
  Yeah, this one involves Meta and their Llama
• 8:13
  AI models. There are reports indicating a pretty
• 8:16
  significant talent drain, you could call it.
• 8:19
  A talent drain? People leaving. A lot of people
• 8:21
  leaving, particularly researchers from the team
• 8:23
  that originally built Llama. Many seem to be
• 8:26
  heading to competitors, especially a company
• 8:28
  called Mistral. Mistral, okay. How significant
• 8:30
  is the string? Well, consider this. Of the 14
• 8:33
  authors on the original llama paper from 2023,
• 8:36
  only three are reportedly still at meta. Only
• 8:38
  three out of 14? Wow, that's substantial. It
• 8:41
  really is. Losing that many foundational members
• 8:43
  can definitely pose a challenge, you know, for
• 8:46
  continuity and future innovation. And Mistral,
• 8:48
  you mentioned, they're snapping up this talent.
• 8:51
  Seems that way. Mistral AI is a startup based
• 8:53
  in Paris. Interestingly, it was co -founded by
• 8:56
  two former Meta researchers who are key architects
• 8:59
  of Lama Guillaume Lampel and Timothee Lacroix.
• 9:02
  Ah, so the founders came from Meta's Lama team.
• 9:05
  Yep. And now several other Meta alumni work there
• 9:08
  building out competing open source AI models.
• 9:11
  This must be happening at a tricky time for Meta's
• 9:13
  AI efforts too. It seems so. There are reports
• 9:16
  their biggest planned AI model Codenamed behemoth
• 9:19
  has been delayed due to some internal concerns
• 9:22
  and their latest public model llama for Apparently
• 9:25
  got a bit of a lukewarm reception from developers
• 9:28
  Some are looking more towards rivals like deep
• 9:31
  seek or Quinn now So maybe losing momentum and
• 9:33
  wasn't there a leadership change too? Yes, Joel
• 9:36
  Pinot who led metas fundamental AI research group
• 9:38
  fair for eight years stepped down recently Robert
• 9:41
  fergus has taken over so leadership change talent
• 9:46
  leaving This really could challenge metas whole
• 9:49
  open source AI strategy, couldn't it? It could
• 9:52
  that original llama paper back in 2023 was a
• 9:55
  big deal It really legitimized these open -weight
• 9:58
  large language models as serious alternatives
• 10:01
  to proprietary ones like opening eyes GPT or
• 10:04
  Google's Paul and they kind of opened the door
• 10:06
  for that whole movement they did they train llama
• 10:09
  on public data focused on making it efficient
• 10:11
  enough to run on like a single GPU, which made
• 10:14
  it accessible. But yeah, that early lead they
• 10:16
  had in open source AI might be slipping. And
• 10:18
  they still don't have a dedicated reasoning model,
• 10:20
  which seems increasingly important for complex
• 10:22
  AI. That's another reported gap, yeah. These
• 10:25
  departures don't help. And these aren't junior
• 10:27
  people leaving either. The average time the 11
• 10:29
  departed authors spent at Meta was over five
• 10:31
  years. So, losing a lot of experience, where
• 10:33
  are they all going besides Mistral? All over,
• 10:36
  really. We mentioned Mistral has quite a few
• 10:38
  Baptiste Rozier, Timothee Lacroix, Marianne Lachaud,
• 10:42
  Thibaut Lavril, Guillaume Lampel. But others
• 10:45
  went to Namon Goyle at Thinking Machines Lab,
• 10:48
  Aurelien Rodriguez to Cohere, Eric Hambro to
• 10:52
  Anthropic, Armand Jolain to Google DeepMind,
• 10:56
  Gaudier Isacard to Microsoft AI, Edward Grave
• 10:59
  to Qtai, another French AI lab. It's a real migration.
• 11:03
  That definitely highlights how fierce the competition
• 11:06
  for top AI talent is right now. This could potentially
• 11:10
  weaken Meta's innovation pipeline down the road.
• 11:12
  It's a concern, for sure. Meanwhile, they are
• 11:15
  still working on Llama 4, apparently trying to
• 11:17
  tune it to address perceived biases, and they're
• 11:20
  also dealing with that copyright lawsuit. Oh
• 11:22
  yeah, the one about training data potentially
• 11:24
  using pirated books. That's the one. A judge
• 11:26
  is looking into whether it actually harmed the
• 11:28
  market for authors. So lots going on, but the
• 11:31
  talent departures are a major storyline. It just
• 11:33
  shows, you know, in this AI race, holding onto
• 11:35
  your key people is absolutely critical. No kidding.
• 11:38
  It's an incredibly competitive landscape. For
• 11:40
  sure. OK, so let's quickly recap the three big
• 11:42
  stories from our deep dive today. First, Salesforce
• 11:45
  making a significant $1 .9 billion move to acquire
• 11:49
  own company, bolstering its cloud data protection
• 11:52
  offerings. Then SpaceX really pushing forward
• 11:55
  with ambitious plans for an uncrewed Starship
• 11:58
  mission to Mars, targeting 2026 despite challenges.
• 12:02
  And finally, Meta facing a notable talent drain
• 12:05
  from its core Llama AI team, with many key researchers
• 12:08
  moving to competitors like Mishraal, highlighting
• 12:11
  that intense battle for AI expertise. Right.
• 12:14
  And all these things, developments in data security,
• 12:16
  the push for space exploration, the shifts in
• 12:18
  AI talent, they all connect back and shape the
• 12:21
  technology landscape we all navigate. Definitely.
• 12:23
  big shifts happen. Absolutely. Well, that's our
• 12:25
  time for today. Remember, we are here to help
• 12:28
  here at Snarful, so reach out on our website
• 12:30
  if you have any more questions. That's a wrap
• 12:35
  on today's top stories. Remember, tech doesn't
• 12:38
  have to be complicated, and we're here to help
• 12:40
  connect the dots. If you've got questions or
• 12:43
  want to learn more, visit us at snarful .com.
• 12:47
  Thanks for listening and we'll catch you next
• 12:49
  time.