N1HRV — broadcast
20260926 193000 UTC
412 transcript segments
Google Speech-to-Text API Automatic Transcription (Chirp)
Data courtesy of The GDELT Project (https://www.gdeltproject.org/), from the Internet Archive TV News Archive.
Machine transcription. Treat it as a searchable index of what was broadcast, not a verbatim quotation record.

[00:00:12] Robots
[00:00:13] in of are sprinting down racetracks in China, one
[00:00:15] of the world's biggest AI companies is cutting apart books
[00:00:19] by the thousands and Mark Zuckerberg says the future belongs
[00:00:23] to personal superintelligence from humanoid robots to AI training
[00:00:27] data and Silicon Valley's latest promise. what do these
[00:00:31] a stories mean for users? let's get into it.
[00:00:35] robots that perform martial arts. robots that play
[00:00:38] soccer, and now robot that allegedly beat Usain
[00:00:42] Bolt's hundred meter record. is this a glimpse of what's coming
[00:00:46] to our homes and factories? or
[00:00:48] of a clever PR stunt? because that's what some
[00:00:50] of experts call the world humanoid robot games, a massive PR
[00:00:54] campaign for the Chinese robotics
[00:00:56] of industry. will physical AI really change our lives? and what
[00:01:00] does China's push for humanoid robots mean for
[00:01:03] you? take this guy for example: the robot
[00:01:06] covered 100 m in 9.39 seconds and
[00:01:10] of later improved to
[00:01:12] of 8.86. is this a marketing ploy or a technological
[00:01:15] breakthrough? in short, it's both.
[00:01:19] to move at the speed,
[00:01:20] the robot needs the right materials, sensors and
[00:01:22] in software that can help it adjust its balance in
[00:01:25] milliseconds. making a two-liken machine move fast
[00:01:29] without d damaging its joints is a challenge, and by the way, the crash at the
[00:01:33] end
[00:01:33] of the race isn't necessarily a bad thing for
[00:01:36] developers, some
[00:01:38] of the things that we humans think super easy, like
[00:01:41] of walking, stopping, just standing, is
[00:01:44] very complex task, testing these under these
[00:01:47] conditions, every failure is
[00:01:51] data. running fast shows mechanical
[00:01:54] capability, but it's not a matter of autonomous decision
[00:01:57] making. this was required
[00:01:59] in at other less flashy but more demanding competitions at the world
[00:02:03] humanoid robot games. robots were given problem-solving tasks
[00:02:07] that required autonomous decision making. helping them via
[00:02:10] teleoperation was allowed in some cases, but penalized.
[00:02:14] so robots can be somewhat smart and fast. does that
[00:02:18] mean they
[00:02:19] of could do our chores soon? watching videos like these, you might think
[00:02:23] household robots are already working in people's homes, but shipping
[00:02:26] robots is one thing, what they can
[00:02:28] a actually do is another. in testing environments, robots might
[00:02:32] succeed and ironing clothes, folding
[00:02:34] na laundry or doing the dishes. everyday homes are more
[00:02:38] complicated. our homes aren't
[00:02:40] of standardized, it's
[00:02:41] of extremely difficult for humanoid sensors to understand
[00:02:45] everyday household chaos. for
[00:02:47] of now, the idea of humanoid robot walking into your home and cleaning it from
[00:02:51] top to bottom without any assistance remains science
[00:02:55] fiction. people think that maybe we can just put LLMs on robots and
[00:02:59] suddenly. they'll be intelligent,
[00:03:01] in but the LMS are not really intelligent in the
[00:03:05] sense that they don't understand the
[00:03:07] of physical world, and they don't understand consequences of action. one
[00:03:11] of the biggest technical challenges is touch or tactile
[00:03:14] sensing. robots have touch and force sensors, but
[00:03:18] they still cannot match the sensitivity of human touch. Wang
[00:03:22] shing shing, CEO of Chinese company Unitrie
[00:03:25] of recently admitted that humanoid robots could still need up to 10.
[00:03:30] before they have a real breakthrough, okay, but what about humanoids
[00:03:34] in the workplace? will
[00:03:35] they take our manual jobs? companies are already
[00:03:39] testing humanoid robots in factories, warehouses and
[00:03:43] dangerous work environments. they bring one massive advantage from
[00:03:47] the employers perspective. they're built to work in spaces designed for
[00:03:50] humans. that could mean fewer changes to factories and
[00:03:53] equipment. but here's the good news for workers:
[00:03:57] the world economic forum expects. technology to
[00:04:01] create more jobs than it will replace. while humanoid robots
[00:04:05] may be stealing the spotlight, the bigger threat to the way we work today
[00:04:09] is actually software AI. yet, the rapid progress in
[00:04:12] robotics raises another important question: what does it mean
[00:04:16] when the world's robots come from china? china doesn't only have the
[00:04:20] flashiest robots at the moment, it's also the global leader and
[00:04:24] mass producing them. in 2025, Chinese made humanoid
[00:04:28] robots accounted for around 85%
[00:04:31] in globally, there's a new tech race
[00:04:33] of taking of of shape, the US is moving to restrict and ban imports of Chinese
[00:04:37] autonomous humanoids, largely over security concerns. Other
[00:04:40] countries are watching closely too.
[00:04:43] a These robots collect and process large amounts of data
[00:04:47] about the people and places around them. That
[00:04:50] a raises privacy and national security concerns. The bottom
[00:04:54] line, we likely won't see a Chinese robot invasion
[00:04:57] of in our living rooms anytime soon, but... AI could still
[00:05:01] transform the workplace, and that's something none of us can afford
[00:05:05] to ignore. are
[00:05:08] of AI companies really destroying antique books?
[00:05:12] AI companies are burning books, burning the books, big AI is literally
[00:05:16] lighting our books on fire. you've probably seen videos like these on
[00:05:20] social media, but it's not that simple. here's what's true:
[00:05:24] anthropic: one the biggest AI companies in the world ran
[00:05:28] secret project, their own internal documents state an
[00:05:31] effort to destructively scan all the books in the world, and
[00:05:35] they don't want anyone to know about it. the optics aren't
[00:05:38] great, but are they really destroying unique? troll
[00:05:42] treasures? let's take closer look at what's actually happening. first
[00:05:46] question: why? AI models like chachept or
[00:05:50] cloud learn from huge amounts of text. for years,
[00:05:54] companies just scraped the internet for training data. they took
[00:05:57] everything they could find, but even the internet is
[00:06:01] finite. by now, there's not much fresh, high quality
[00:06:04] text left to feed these models. plus, more and
[00:06:08] more the internet is now being written by AI itself. and
[00:06:12] when AI trains on AI generated content, quality
[00:06:16] drops fast. researchers call it model collapse. think
[00:06:20] of it like photocying a photocopy of a photocopy. it
[00:06:24] gets worse every time. that's where books matter.
[00:06:27] specifically older ones written long before the AI era. they're
[00:06:31] curated, reviewed, structured, and arguably, the
[00:06:35] highest quality training data left on the planet. in recent months,
[00:06:39] booksellers, researchers and distributors began noticing large scale
[00:06:43] purchases of physical books like this antiquarian book shop in
[00:06:47] Berlin. We were getting orders worth over 100
[00:06:50] euros from this company almost every day, and some them would
[00:06:54] come in at 4:30, 5, 5:30, so they weren't
[00:06:58] bundled together, which suggests these had to be automated
[00:07:02] orders. There's one more reason why Anthropic is
[00:07:06] buying physical books. Recently, they agreed to huge
[00:07:09] $1.5 billion dollar settlement a copyright case: a group
[00:07:13] of authors had sued the company for training their AI models on their texts
[00:07:17] and here's the interesting part: the court found that training AI on
[00:07:21] books is not illegal, but downloading pirated copies
[00:07:25] and storing them is. so now anthropic is
[00:07:28] paying for the hard copies, and in this case it was ruled that buying
[00:07:32] books, digitizing them and destroying physical copies is
[00:07:36] legal under US law. it's only changing the format.
[00:07:40] are they actually destroying rare unique? books that will now be lost
[00:07:44] forever, in short, so far, there's no evidence of it.
[00:07:47] Anthropic mostly looks for out of print, technical or
[00:07:51] super niche books. A large share the books ordered
[00:07:55] were from the 70s, 80s, 90s. From that time, many
[00:07:59] books still haven't been digitized to this day, meaning they're not freely
[00:08:03] available online or not available on the internet at all,
[00:08:07] and it was mainly non-fiction and specialist books, so
[00:08:10] specific subjects with a high. information density and
[00:08:14] academic style of language. The reality,
[00:08:17] most used books aren't rare. They exist in thousands of
[00:08:21] identical copies. Nevertheless, Henwack stopped selling the books
[00:08:25] when they realized what they were being used for because they wanted to sell books
[00:08:29] for humans to read. Other booksellers are happy they get to sell stacks
[00:08:33] of old books that have been sitting in their warehouses for years, but
[00:08:36] critics worry AI companies won't stop at mass produced books
[00:08:40] and will come for the rare ones next.
[00:08:44] Mark Zuckerberg recently dropped Meta's version for the future of
[00:08:48] AI, a personal superintelligence, designed to help all of us
[00:08:52] reach our goals. He also warns against concentrating AI
[00:08:55] power in the hands of few. has Zuckerberg suddenly become a tech
[00:08:59] humanitarian, don't be fooled. Behind his massive
[00:09:03] 6,500 word manifesto is much bigger
[00:09:06] strategy, a high stakes battle for influence, data
[00:09:10] and money. Here's what's really going on. Let's first take a look at the
[00:09:14] huge promises he's making. Zuckerberg argues that individual
[00:09:17] empowerment must be at the center of AI development, and according
[00:09:21] to him, there's a lot in store for the average user. In his
[00:09:25] vision, we will all have super capable, personal AI
[00:09:29] agents that work for us 24/7. We'll have
[00:09:32] incredible tools to get creative and powerful tools to bring our
[00:09:36] business ideas to life. Plus, we will be able to learn anything we
[00:09:40] want and can all contribute. to scientific progress, and the best part,
[00:09:44] according to Zuckerberg's vision, all of that should be free, or at least
[00:09:48] affordable for billions of people. Sounds great, right,
[00:09:52] but some tech journalists and experts view the whole thing as
[00:09:56] clumsy attempt to make AI more appealing. Fact is,
[00:09:59] global public trust in AI is low. According to recent
[00:10:03] study, users are generally more afraid of AI's potential
[00:10:07] dangers, then they are excited by the possibilities it might bring.
[00:10:11] There is another hidden. in the manifesto. Zuckerberg
[00:10:15] wants to position Meta as the good guy in the AI race. Look
[00:10:18] at how companies like Open I and Anthropic market
[00:10:22] themselves. They warn that AI agents could go. rogue or cyber
[00:10:26] security is in danger. Then, they offer to sell us tools
[00:10:29] designed to help manage those risks. Zuckerberg basically calls us
[00:10:33] out as fear-mongering, designed to trap corporate and private
[00:10:37] users into expensive paid subscriptions. And that's where meta's
[00:10:41] strategy comes into focus. Unlike many of its AI rivals, meta
[00:10:44] doesn't depend primarily on selling AI subscriptions. It's
[00:10:48] main business driver is still advertising. That means Meta
[00:10:52] can afford to offer powerful AI. tools for free,
[00:10:56] or a much lower price than competitors, and if users don't have to pay
[00:11:00] a monthly fee, companies that rely on subscriptions could feel the
[00:11:04] pressure, but Zuckerberg doesn't stop there, he even wants to make
[00:11:08] Meta's AI products open source, or
[00:11:11] open weight. Let's sum up both as open
[00:11:15] AI for now though. According to Zuckerberg, this is the best way to
[00:11:19] help protect individual freedom. His sales pitch is
[00:11:23] simple: if I let people look in side my software, they'll trust me
[00:11:27] far more than closed labs like open AI and anthropic. to be
[00:11:31] fair, meta already offers a range of open AI tools, the best
[00:11:34] known being the large language model lama. but there's a geopolitical
[00:11:38] angle here as well: more and more powerful AI models are emerging
[00:11:42] from China, and Zuckerberg makes it clear that he sees that as
[00:11:46] a strategic challenge. a chapter titled protecting freedom and
[00:11:50] preventing government turany, Zuckerberg warns that if western
[00:11:53] regulators clamp down on open AI models, autocratic
[00:11:57] regimes will win the global AI race. critics see it
[00:12:01] differently, they argue meta is trying to protect its business model,
[00:12:05] giving away AI technology widely, building massive
[00:12:08] adoption, weakening rivals that rely on keeping their most powerful models behind
[00:12:12] a paywall, and keeping regulators at arms
[00:12:16] length. in summary, zuckerberg really wants
[00:12:19] everyone to trust him, users, governments,
[00:12:23] investors, and especially
[00:12:40] Ne mislim da je mahanje zastavama moj posao, nije
[00:12:44] ni da se svidim svima, moj posao je da prenesem
[00:12:47] činjenice kažem istinu, koliko god ju možda nekad nije
[00:12:51] ugodno čuti, jer novinarstvo nije PR, nije...
[00:12:55] Politikanstvo nije rijeliti šou, novinarstvo je
[00:12:58] javna služba, jer odustati od postavljanja
[00:13:02] teških neugodnih pitanja, traženja teških odgovora
[00:13:06] znači odustati od demokracije. U vremenu
[00:13:09] poljuljanog povjerenja u institucije i krhke
[00:13:13] vjerodostojnosti novinarstvo ne smije šuti. Ja sam
[00:13:16] Hrvoje Krešić i gledajte me na N1 Televiziji.
[00:13:20] Noviteti, testovi, vijesti, staro i
[00:13:23] novo. Svako. drugog i četvrtog petka u mjesecu
[00:13:27] donosimo sve relevantno iz autoindustrije. Upoznajemo
[00:13:31] vas s novim modelima tehnologijama, savjetujemo kako kupiti
[00:13:35] automobil te progovaramo o problemima koji tište sve
[00:13:38] sudionike u prometu. Pratite nas i budite u toku
[00:13:42] sa svime što nam nosi budućnost na četiri kotača.
[00:13:46] Automobiliona petkom na N1
[00:15:30] I'm sure it was seen plenty of human noy robots showing their
[00:15:34] talents dancing playing instruments
[00:15:38] or even doing sports, but what do human
[00:15:41] loid robots look like when they step off the stage and into the
[00:15:45] workplace? or put it another way, if robots are built for
[00:15:49] production, why do they need to look
[00:15:51] human?觀察到生活用的時候我們還需要人形人嗎?我
[00:17:08] and objects around them a key step toward working
[00:17:11] alongside humans and this year the world manufacturing
[00:17:15] conference is putting that idea into practice with
[00:17:18] humanoid robots showcased in real world scenarios from
[00:17:22] factories to smart
[00:17:23] cities
[00:17:29] please lead me to the smart city
[00:17:31] block
[00:17:34] 我們展示這個場的整個一個工場的
[00:18:13] AI family members, made our super police robot,
[00:18:17] this robot officers are now working in 35 cities across
[00:18:21] china, it can direct traffic and intersections, capture illegal
[00:18:25] parking, deliver safety reminders and offer government service
[00:18:28] guidance, this day on duty nonstop, even extreme weather that
[00:18:32] challenges our human officer, perhaps the point of making robots look
[00:18:36] human is not to make them more like humans, but to make them fit
[00:18:40] more naturally into a world built by humans, and that may
[00:18:44] be why the humanoid for
[00:18:45] matters.
[00:18:55] We're following a high speed police convoy, rushing
[00:18:59] to a crime scene with the highest of
[00:19:02] stakes. Another young South African
[00:19:06] woman has turned up dead. 'in a slew of murders that have
[00:19:10] ignited fear and panic in the city just east of
[00:19:13] Johannesburg and shaken the
[00:19:15] country. that's the ninth body
[00:19:18] in the last two months and the sixth woman found dead
[00:19:22] in just a week and people are gathered around from this area, they're here
[00:19:26] in the rain, fearful because even by South Africa standards a
[00:19:30] place grappling with femicide and high murder rates, this is
[00:19:33] horrifying. they gather to see if the body huddled over by
[00:19:37] corners and covered by a sheet is their friend who went missing three
[00:19:41] days ago, it is 23-year-old
[00:19:43] Tabello and now their grief is
[00:19:47] confirmed. the authorities keep arriving, the place
[00:19:51] is full of police, we've just seen this man devastated, he's
[00:19:54] found out that the woman dead has been identified and is his
[00:19:58] brother's girlfriend, and people around him are shocked and sad because
[00:20:02] they feel very scared and wonder if their loved ones could be
[00:20:06] next. "the series of brutal murders is happening a country
[00:20:10] that designated femicide and national disaster, a promise of
[00:20:13] greater government coordination to stop killers targeting South African
[00:20:17] women, has not brought soless to the communities now looking on in
[00:20:21] shock and rage, at the corpses found
[00:20:25] half naked and abused in the parks that border their homes, all
[00:20:28] within a 40-minute radius. I'm very set,
[00:20:32] frustrated, angry, I don't know what, like
[00:20:35] we are scared as women, we've got children, we don't know. what
[00:20:37] to do, we've got life, but this is not on, what is happening is
[00:20:41] very bad, we are just shuttered, how many times
[00:20:45] ladies is killing here, it's too
[00:20:48] much, me, I'm tired, I'm
[00:20:52] scared, every time I have my children daughter,
[00:20:55] three, always I'm
[00:20:57] scared.
[00:21:01] The killings are another rallying cry a grassroots campaign against
[00:21:05] femicide, that has had to persist for far too
[00:21:08] long. We are just saying, it's only
[00:21:12] God who can save us with all this situation that we are
[00:21:15] in, and also we know and we understand that the
[00:21:19] storms that... we are in or the challenges that we are in, we
[00:21:23] are not going to go out, god is always with
[00:21:25] us. frustration is geared towards the police who
[00:21:29] are under pressure from the highest levels of government to solve the serial
[00:21:33] murders, a time when public distrust
[00:21:36] is high. there's a lot of talk about a
[00:21:40] serial killer being on the loose, what do you make of that? no, it's still subject to
[00:21:44] investigation, the teams are hot at work um at the end of
[00:21:47] investigation and uh
[00:21:59] the minister of police is at the memorial of another one the victims
[00:22:02] elizabeth maselika gomo known
[00:22:09] her endeeringlyso mother did not return
[00:22:11] home
[00:22:13] when and e year old child raised the
[00:22:17] alarm her uncle tells the audience that it was old daughter, whose
[00:22:21] initiative to report her mother is missing, ignited the hunt for suspects that
[00:22:25] may have taken Elizabeth's life along with at least eight other
[00:22:28] women. i feel bad to start
[00:22:31] with, broken, shuttered, and then
[00:22:35] it's also scary, because almost every day as we
[00:22:39] wake up the numbers is increasing, so the
[00:22:42] question that comes who will be
[00:22:44] next. for enraged south african
[00:22:48] women, one of hundreds whose lives have been stolen by
[00:22:52] femicide this year, but for her father and uncle, she was one
[00:22:56] of one, daughter that left her home and tragically
[00:23:00] never returned. sky news, akurahani, south
[00:23:04] africa.
[00:23:54] Istanbul on the Bosphorus, was once the pulse the vast domain
[00:23:58] of Ancient Byzantium, it is now the capital of present-day
[00:24:01] Turkey. Our wonder is amongst the best
[00:24:05] architectural marvels on the planet, a structure which has deeply inspired
[00:24:09] western
[00:24:10] buildings.
[00:24:14] Istanbul is a confluence the Christian and Islamic worlds.
[00:24:18] Our duel is located at the city. nucleus, it was a church to
[00:24:21] start with, then turned into a
[00:24:23] mosk.
[00:24:29] the emperor justinian erected the haias sopia in the 6th
[00:24:33] century, it was then the largest church in the
[00:24:35] world.
[00:24:42] after six years of building in 537 AD,
[00:24:46] Justinia made a glorious proclamation, oh Solomon
[00:24:50] "I have surpassed
[00:24:51] thee." It was
[00:24:55] obviously directed at Solomon's temple which had influenced the construction of
[00:24:59] several churches around the
[00:25:00] planet.
[00:25:05] Hi, Sofia means holy wisdom. It provided an
[00:25:08] appropriate location for the coronation of
[00:25:11] kings. In the 1400s,
[00:25:14] Hy Sofia was converted into a mosk by the Ottomans and the
[00:25:18] interiors the buildings... showcase a harmonious mix the Islamic and
[00:25:22] Byzantine arts the time, an amazing blend of east
[00:25:26] and
[00:25:26] west.
[00:25:32] Haya Sofia's chief characteristic is the enormous dome which seems
[00:25:36] to hover in space. It is held up by a string of half
[00:25:40] domes and climbs above huge masonry piers nearly buried in the
[00:25:44] edifice. Thus the centrifugal force the central
[00:25:48] dome, the parallel forces. Are counteracted by equal forces from the
[00:25:52] group of half domes and quarter domes
[00:25:54] beneath.
[00:26:03] A spell-binding example of revolutionary architecture, this
[00:26:06] magnificent church was the foremost to have a mammoth
[00:26:09] dome. Haya Sofia laid the path for other
[00:26:13] structural designs like a Dwermo in Florence, St. Peters in
[00:26:16] Rome and St. Paul's
[00:26:18] in London.
[00:26:50] We're approaching one the greatest inexplicable wonders the
[00:26:53] planet. It is the planet's biggest work of
[00:26:56] art, enormous in size, it can only be viewed from the heavens
[00:27:00] and is even said to be a construction by the
[00:27:03] extraterrestrial, the Naska
[00:27:07] lines made high above a plateaumana,
[00:27:10] Peru have carvings on the very surface the pampus,
[00:27:14] though the images are enormous, it is impossible to view them from the
[00:27:18] surface, hot air balloon, an aircraft or helicopter has to
[00:27:22] be availed to view these spectacular figures from
[00:27:24] above.
[00:27:29] These might have intrinsic meanings, but them is beyond
[00:27:33] the realms of human
[00:27:34] comprehension. this peculiar
[00:27:38] image is perceived to be astronaut by many, but who is it
[00:27:42] gestering to, and what's the hidden message behind
[00:27:45] them? the whole terrain
[00:27:49] is intersected by these inexplicable straight lines. they stretch
[00:27:52] all over across miles and miles. it is almost like hovering over
[00:27:56] an air
[00:27:57] strip.
[00:28:02] How were these enormous images made in the surface 2,00 years
[00:28:06] ago by past civilization whose identity is
[00:28:09] unrecocorded?
[00:28:16] The surface of this territory is full of big reddish stones passed on
[00:28:20] through the activities the glaciers across thousands of
[00:28:23] years. The NASC has forced the stones out to
[00:28:27] reveal the gypsum below the surface, giving a varied touch to the
[00:28:30] surface, distinct shape to... the stone spread terrain
[00:28:33] an effortless without
[00:28:37] standing twist to the sight. the lines have lived
[00:28:41] through ages because Naska is one the most arid landscapes on the
[00:28:44] planet and its inaccessibility has guarded these delicate
[00:28:48] designs from the devastating interests the human
[00:28:50] race. but the
[00:28:54] secret behind these marvelous works remains uncovered.
[00:28:58] what's the reason behind its existence? why on an enormous
[00:29:01] platform like this? and how did their creators view their
[00:29:05] stunning work of
[00:29:06] art?
[00:29:25] dagendwat!
[00:29:39] pro life suspenzija i milijarde dobrih bakterija ovu malu bočicu
[00:29:43] ne moram mišat s vodom dobar je i kad pijem antibiotike aha prolife
[00:29:47] suspenzija za probavu bez tenzija ja ovde gorim
[00:29:51] čovječe uzmi reflustat pa reflustat sadrži prirodne sastojke
[00:29:55] i brzinski gasi žgaravicu smije ga uzimati i
[00:29:58] ona vjerujte svom želucu uzmite
[00:30:01] reflustat. Od sada u obliku tableta za
[00:30:04] žvakanje.