Latest / Connecting the Dots / Grammarly’s AI Leap, China’s Robot Games, and SoftBank’s Stargate Move

Transcript

Grammarly’s AI Leap, China’s Robot Games, and SoftBank’s Stargate Move · Connecting the Dots

• 0:00
  Today's tech news moves fast, but understanding
• 0:03
  it shouldn't be hard. That's why Snarfle 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 to Connecting the Dots from Snarfl
• 0:17
  Solutions Group. I'm Alex and I'm Morgan. Great
• 0:20
  to be with you today, Monday, August 18th, 2025.
• 0:23
  All right, let's get everyone up to speed. First,
• 0:25
  the weather picture nationally. We've got major
• 0:28
  hurricane Aaron tracking just north of the lower
• 0:31
  Bahamas. Yeah, that's bringing some dangerous
• 0:33
  surf and rip currents to. pretty much the whole
• 0:35
  Atlantic coastline. Outer bands might even clip
• 0:38
  the outer banks later this week. And it's going
• 0:40
  to be hot, extremely hot actually, today and
• 0:42
  Tuesday from the Southern Plains right into the
• 0:45
  central and lower Mississippi River Valley. Oof.
• 0:48
  Stay cool out there. Yeah, but closer to home
• 0:51
  here in Sacramento much nicer sunny today high
• 0:54
  around 87 degrees Fahrenheit pretty pleasant
• 0:56
  actually can't complain about that Okay, quick
• 0:59
  market snapshot looking back at Friday August
• 1:01
  15th close right the Dow Jones finished at forty
• 1:04
  four thousand nine hundred and forty six point
• 1:06
  one tell that was down about 213 points for the
• 1:09
  day minus point four seven percent But still
• 1:12
  up for the week, right? Yeah, up 1 .72 % for
• 1:15
  the week. So not too bad overall. The S &P 500
• 1:17
  closed Friday at 6449 .80. OK, and that was down
• 1:22
  a bit, too. Down just 27 points or so, minus
• 1:26
  0 .43 % for the day. But like the Dow up for
• 1:30
  the week just under 1 .94%. Got it. And how's
• 1:33
  Bitcoin looking today, Monday the 18th? Bitcoin
• 1:36
  closing at 115 ,662 .73. It's down about $1 ,800
• 1:41
  today or one and a half percent. Okay. And actually
• 1:44
  down about two and a half percent for the week
• 1:46
  as well. So a bit of a dip there. Interesting.
• 1:48
  All right. Before we jump into tech, some important
• 1:50
  news. Yes. The Giants. Huge win yesterday, Sunday
• 1:54
  the 17th. Tell us about it. They beat the Tampa
• 1:57
  Bay Rays 7 -1. Fantastic game. Oh, that's great
• 2:00
  news. Who pitched? Logan Webb got the win. Solid
• 2:03
  performance. And Drew Gilbert had a great day
• 2:05
  at the plate. Went 2 -4 with a home run. Nice.
• 2:08
  Love to see that. A good win always makes Monday
• 2:11
  feel a bit better, doesn't it? It absolutely
• 2:13
  does. Always happy when the Giants pull out a
• 2:15
  win like that. Definitely. Okay, so feeling good
• 2:17
  about that. Let's connect some dots on today's
• 2:20
  top tech stories from the Morningstar. Perfect.
• 2:23
  First up, a company many people use... uh, Grammarly.
• 2:27
  They're making some really big moves with what
• 2:29
  they're calling AI agents. Yeah, and not just
• 2:31
  agents, but a whole new writing environment called
• 2:34
  Docs, described as an AI native writing surface.
• 2:38
  AI native. So built around AI from the ground
• 2:41
  up, not just adding features. Exactly. They're
• 2:44
  launching these specialized AI agents designed
• 2:46
  to help with specific writing challenges. And
• 2:49
  this isn't just for you know premium users it's
• 2:52
  for grammarly free and pro okay so wider access
• 2:55
  what kind of agents are we talking about here
• 2:56
  for students maybe yeah definitely focusing there
• 2:59
  there's an ai grader which is pretty interesting
• 3:01
  it actually tries to predict your grade based
• 3:03
  on uh instructor info and gives feedback before
• 3:06
  you submit wow like a pre -check kind of and
• 3:09
  then there's a citation finder this one sounds
• 3:11
  like a lifesaver automatically generates formatted
• 3:14
  citations and scours the web for evidence to
• 3:16
  back things up. Okay, that alone is huge. No
• 3:18
  more wrestling with APA or MLA at 2 a .m. Right,
• 3:22
  and there's more for students. A reader reactions
• 3:24
  agent to predict reader questions, the standard
• 3:27
  proofreader, a paraphrase agent, even an expert
• 3:30
  review for topic -specific feedback. It sounds
• 3:33
  less like a tool and more like a, well, writing
• 3:35
  partner. That seems to be the goal. And for educators,
• 3:39
  specifically pro -users, they're adding a plagiarism
• 3:41
  checker, scanning huge databases, and an AI detector.
• 3:45
  Ah, the AI detection piece. That's always tricky.
• 3:48
  How are they positioning that? Well, their VP,
• 3:50
  Luke Banghan, was pretty upfront, said AI detection
• 3:53
  can be hit or miss, which is, you know, honest.
• 3:56
  Yeah. But he stressed their tool is tuned for
• 3:59
  accuracy, and it's meant more as a window to
• 4:01
  students before they submit. kind of a guide,
• 4:04
  not just an enforcement hammer. OK, so focusing
• 4:06
  on the learning process, that makes sense. So
• 4:07
  why does all this matter for, you know, real
• 4:10
  people day to day? Well, for students, Grammarly's
• 4:14
  head of education, Jenny Maxwell, put it nicely.
• 4:17
  She said it enhances capabilities without undermining
• 4:20
  learning. It's about building AI literacy. So
• 4:23
  teaching them to work with AI, not just rely
• 4:25
  on it to cheat, preparing them for the workforce
• 4:28
  where AI is, well, everywhere. Exactly. It bridges
• 4:31
  that gap. And for educators, it helps maintain
• 4:34
  academic integrity while still guiding students.
• 4:37
  For professionals, think about it. Like if I'm
• 4:39
  a marketing director. Right. Use that reader
• 4:42
  reactions agent to see what questions your audience
• 4:44
  might have about a new campaign before you launch
• 4:46
  it or just saving time on what they call rote
• 4:49
  busy work. That term resonates. Formatting, basic
• 4:53
  proofing, if AI can handle that reliably. It
• 4:57
  frees you up for the bigger picture stuff, the
• 4:59
  strategic thinking like you said about citations
• 5:01
  earlier. Imagine the hours saved across the board.
• 5:04
  It's a really interesting balance they're trying
• 5:05
  to strike, isn't it? AI for writing and AI for
• 5:08
  detecting AI writing. It definitely shows their
• 5:10
  bigger strategy. They want to be a platform built
• 5:13
  around these proactive agents moving way beyond
• 5:16
  just fixing grammar. And those recent acquisitions
• 5:19
  CODA, Superhuman, plus that massive, what, billion
• 5:22
  dollar funding round? Yep. Yeah. It all points
• 5:24
  to a serious commitment to this agent -centric
• 5:27
  future for them. It's a big bet. A very big bet.
• 5:30
  OK, so from AI writing assistants, let's shift
• 5:33
  gears to robots that are, well, maybe not quite
• 5:37
  writing essays yet, but trying to play soccer.
• 5:39
  Huh. Yeah, let's talk about the humanoid robot
• 5:42
  games held recently in Beijing. Sounds intriguing.
• 5:44
  What was the takeaway? Were they amazing? Were
• 5:46
  they terrible? Uh, kind of both. It was a big
• 5:49
  event. Three days, over 280 teams, 16 countries.
• 5:53
  It definitely showed some impressive steps forward
• 5:55
  in robotics. Like what, what cool things could
• 5:57
  they do? Well, we saw robots doing back flips.
• 6:01
  navigating obstacle courses, pretty complex stuff,
• 6:04
  and there were fully autonomous three -on -three
• 6:07
  soccer matches using these booster robotics T1
• 6:10
  robots. Autonomous soccer, so no humans controlling
• 6:13
  them. Nope, all AI. They could identify the ball,
• 6:15
  move around the field, even supposedly stand
• 6:17
  back up after falling, which happened a lot.
• 6:20
  Okay, I was gonna ask about the falling. Oh yeah,
• 6:22
  but some were impressive. One robot from Unitree
• 6:24
  Robotics apparently sprinted faster than many
• 6:27
  amateur humans, but... and this is a big but
• 6:29
  during a track event it actually plowed right
• 6:32
  into a staff member and knocked him over oh my
• 6:35
  gosh is he okay that sounds not good no details
• 6:39
  on injuries but it's a vivid example right speed
• 6:42
  and agility are improving but control safety
• 6:45
  still a ways to go in unpredictable environments
• 6:48
  definitely highlights the safety challenge So
• 6:51
  besides the occasional tackling of staff, what
• 6:53
  were the limitations? You mentioned falling.
• 6:55
  Yeah, lots of falling. The reports describe child
• 6:58
  -sized robots in the soccer games tripping like
• 7:00
  dominoes. And one goalkeeper just sort of watched
• 7:05
  goals go by passively. Not exactly John Luigi
• 7:08
  Buffon up there. Not quite. And the best visual
• 7:10
  robots being carried off the field on stretchers
• 7:13
  by staff after falling. even though they were
• 7:16
  designed to get back up themselves. Okay, that's
• 7:18
  actually kind of funny. A bit humbling for the
• 7:19
  robots. It adds that touch of relatable humor,
• 7:21
  doesn't it? Shows the gap between design and
• 7:24
  messy reality. So why do these games matter beyond
• 7:27
  the entertainment value of falling robots? Well,
• 7:31
  first, it's a crucial real -world testing ground.
• 7:34
  It pushes AI algorithms, hardware, software integration
• 7:37
  way faster than just lab work. It's like rapid
• 7:40
  fire evolution through competition. Stress testing
• 7:43
  in public, basically. Exactly. And second, it
• 7:46
  brings up that safety and trust issue you flagged.
• 7:49
  As robots become more common, we need to trust
• 7:51
  them. Right. The CEO of Booster Robotics, Cheng
• 7:54
  Hao, mentioned building trust as a key goal for
• 7:57
  human -robot interaction. You don't want your
• 7:59
  warehouse bot running you over. Precisely. That
• 8:02
  sprinting incident really underscores the challenge
• 8:04
  of putting powerful, fast robots safely into
• 8:07
  places with squishy humans. It really manages
• 8:11
  expectations too, doesn't it? Yeah. We hear so
• 8:13
  much about AI's power, but seeing the robot stumble
• 8:16
  reminds us it's a journey. Still lots of work
• 8:18
  to do. Absolute powerful, yes. Perfect, not yet.
• 8:21
  It's that human process of iteration and learning
• 8:24
  just applied to machines. Okay, speaking of big
• 8:26
  moves and major tech players, our next story
• 8:29
  involves a project with a truly astronomical
• 8:32
  budget. Yeah, we're talking about SoftBank's
• 8:34
  Stargate AI project. And it seems to have finally
• 8:37
  found a major US hub. Where's that? In Ohio,
• 8:41
  actually. SoftBank has acquired Foxconn's former
• 8:44
  electric vehicle plant in Lordstown. The old
• 8:47
  GM plant. That place has had quite a history.
• 8:49
  So SoftBank bought it for Stargate. Yes. A critical
• 8:53
  step for their, get this, $500 billion AI data
• 8:57
  center plan. Half a trillion dollars. Wow. Okay.
• 9:01
  And Foxconn is just out. No, actually, Foxconn
• 9:05
  will keep operating the plant, but they're pivoting.
• 9:07
  It's now a joint venture with SoftBank to manufacture
• 9:09
  AI servers in that plant. Interesting. So repurposing
• 9:13
  the EV facility for the AI boom makes sense.
• 9:16
  But wasn't Stargate having some trouble getting
• 9:18
  off the ground? Big trouble. It launched with
• 9:20
  huge fanfare. You know, White House event back
• 9:22
  in January 25. Masayoshi Sun talking $500 billion,
• 9:25
  $100 billion in immediate investment. Partners
• 9:27
  like OpenAI, Oracle. It sounded massive. Well,
• 9:30
  it's meant to be. But then reality hit. Six months
• 9:32
  went by and basically nothing happened. No data
• 9:34
  center deals finalized. What went wrong? Reports
• 9:37
  pointed to major disagreements between the partners,
• 9:39
  especially about where to build these things.
• 9:41
  Site locations became a huge sticking point.
• 9:43
  Politics and logistics getting in the way of
• 9:45
  a half trillion dollar plan. Seems like it. Public
• 9:48
  skepticism grew. Oracle CEO Safra Katz even said
• 9:52
  pretty bluntly, Stargate is not formed yet, which
• 9:55
  is not what you want to hear. No kidding. So
• 9:58
  this Ohio deal. Is this the fix? A restart? It's
• 10:02
  being framed exactly like that. A decisive reboot.
• 10:05
  Foxconn's chairman, Young Lou, basically said,
• 10:07
  this solves the big hurdles, power, venue, timing.
• 10:10
  It gets the project moving. Okay, so why should
• 10:13
  our listeners care about some giant companies
• 10:14
  figuring out where to build data centers in Ohio?
• 10:17
  Well, first, the sheer scale. $500 billion over
• 10:20
  five years. The promise of potentially over 100
• 10:23
  ,000 US jobs, that's a massive economic ripple
• 10:26
  effect. Huge potential impact there. And second,
• 10:29
  these data centers, they are the literal physical
• 10:32
  backbone for the future of AI. Think about the
• 10:35
  power needed for models like OpenAI's GPT -5.
• 10:38
  It has to run somewhere. Right, all the AI magic
• 10:40
  we hear about needs immense infrastructure behind
• 10:42
  the scenes. These are the engine rooms. Exactly,
• 10:44
  the silent enablers. And this move also highlights
• 10:48
  the global AI race. While Stargate US was stalled,
• 10:52
  competitors weren't waiting. Like who? Meta is
• 10:54
  spending hundreds of billions on their own data
• 10:57
  centers. OpenAI made its own deals with Oracle
• 10:59
  and CoreWeave for infrastructure. They even launched
• 11:02
  a Stargate Norway, a gigafactory for AI processing.
• 11:06
  So the pressure was really on SoftBank to get
• 11:08
  the US part moving? Immense pressure. It just
• 11:11
  shows even with visionary plans and massive funding,
• 11:14
  these huge partnerships face really complex technical
• 11:17
  construction and just... plain human consensus
• 11:20
  challenges. It's a good reminder that building
• 11:22
  the future isn't easy, even when you have half
• 11:24
  a trillion dollars earmarked for it. Execution
• 11:26
  matters. Absolutely. This Ohio deal is a very
• 11:29
  tangible step forward after a lot of uncertainty,
• 11:31
  a big piece falling into place. OK, so let's
• 11:33
  quickly recap the dots we connected today. We
• 11:35
  started with Grammarly and their new AI agents,
• 11:39
  really pushing beyond proofreading to become
• 11:40
  more like an AI writing partner, aiming to build
• 11:43
  AI literacy for students and efficiency for professionals.
• 11:46
  Then we dove into the humanoid robot games in
• 11:49
  Beijing, saw some amazing robotic progress, but
• 11:52
  also some very humbling, very real limitations
• 11:55
  and safety concerns. Progress isn't always graceful.
• 11:59
  Definitely not. And finally, Softbank's massive
• 12:02
  Stargate AI project, finding a crucial home in
• 12:05
  Ohio, overcoming some big hurdles to start building
• 12:08
  that essential infrastructure for the future
• 12:10
  of AI amid intense global competition. Three
• 12:13
  very different stories, but all showing how AI
• 12:16
  and advanced tech are weaving into education,
• 12:19
  physical reality, and the fundamental infrastructure
• 12:21
  of our world. It's happening fast. It really
• 12:24
  is. Constantly evolving, sometimes messy, but
• 12:27
  undeniably shaping what's next. Makes you wonder,
• 12:30
  doesn't it, how these tools will continue to
• 12:31
  change what we value as uniquely human skills?
• 12:34
  That's a great thought to leave folks with. And
• 12:36
  that's all the time we have for today. We're
• 12:38
  here to help at Snarful, so reach out on our
• 12:39
  website if you have any more questions. That's
• 12:45
  a wrap on today's top stories. Remember, tech
• 12:48
  doesn't have to be complicated, and we're here
• 12:50
  to help connect the dots. If you've got questions
• 12:53
  or want to learn more, visit us at snarful .com.
• 12:56
  Thanks for listening, and we'll catch you next
• 12:58
  time.