Author: dora

iPhone SE4 Info Release

MacRumors recently revealed the latest news about the iPhone SE4, basically confirming several important hardware information:
6.1-inch OLED display
48-megapixel primary lens
USB port
Operator buttons
Compared with the previous three generations of SE models, the hardware configuration of the iPhone SE4 can be said to have some with the times, and even the iPhone 15 are not used in the operation buttons, Apple is very generous to let the SE first to follow. Of course, for fans of smaller phones or LCD screen lovers, the changes to the iPhone SE4 are still very unfortunate.

Nowadays, Apple has made quite a few ‘compromises’ to make the iPhone SE4 sell better. But there are still a lot of unanswered questions waiting to be explored as to whether or not such changes will bring the SE back to the top.

According to the timeline of the first three generations of iPhone SE and the fit of the body design, Phone SE4 was originally supposed to debut early next year, following the shape of the iPhone XR. Perhaps the change in sales of the iPhone 14 series hit Apple’s confidence, the iPhone SE4 project has been delayed, the design program has been changed again and again, until recently there was news of a restart, and determined the initial version of the program.

The design scheme of the iPhone SE4 is quite confusing, and it is not a direct ‘use’ of the mold of the previous generation of the base model iPhone like the previous generations. Sources say that the iPhone SE4 has a 6.1-inch OLED display on the front, a bangs screen shape, and the biometric solution is FaceID. the bezel is the same right-angled bezel as the iPhone 14/Plus, but the mute switch has been replaced with an action button.

Of course, subject to the EU Act, the iPhone SE4 also uses the USB-C interface, but the rate should still be USB 2.0.Image part, the iPhone SE4 continues the traditional single camera program, the sensor is the same iPhone 15 48-megapixel lens, providing computational photography twofold crop. Theoretically, the iPhone 15 supports up to 24-megapixel imaging mode, and the iPhone SE4 with the same sensor should also support it, but it’s hard to say whether Apple will reduce the configuration on this.

The core configuration is very controversial. At present, there are still questions about whether the iPhone SE4 will be equipped with an A16 Bionic chip or an A17 chip, after all, Apple has used state-of-the-art processors in building the first three generations of SE models. However, in fact, there shouldn’t be a huge difference in performance between the A16 Bionic and the A17, and whichever chip is used, it’s going to be quite a step up from its predecessor.

It’s worth noting that the overall iPhone SE4 will also be more expensive than its predecessor due to its overall upgrade. The latest news reveals that the starting price of the new iPhone SE may be set at $499, which is about Rs. 3,644. With no ‘soul’ home button, no more compact body, and even a slightly higher starting price, I wonder if iPhone SE loyal fans will still pay for this long-delayed fourth-generation model?

In 2016, Apple released the first-generation iPhone SE, a small-sized model with a 4-inch retina display but the same A9 processor as the iPhone 6S/Plus, which has become a god machine for many “fruit fans”. Market research organization IDC released smartphone sales statistics report shows that the cumulative sales of the first generation of iPhone SE reached 30 million units. It can be said that the first-generation iPhone SE can definitely be called a model with both good reputation and sales.

As for why the first-generation SE can be so successful? There’s really no other reason than one: value for money. The overall performance of this generation of iPhone SE does not lag behind the iPhone 7 series released in the same year, and its classic right-angle bezel styling is also deeply rooted in people’s hearts, and Apple even thoughtfully upgraded its camera to 12 megapixels to support 4K video recording. Comprehensive strength can play, the price is not high, become a generation of god machine is not wrong.

After four years, Apple launched the second generation of iPhone SE, which adopts the same design as iPhone 8, and the processor is equipped with the strongest A13 Bionic chip at the moment. However, since Apple has entered the era of the bangs screen and the entire iPhone 12 series released later in the same year was upgraded to the 5G network, the second-generation iPhone SE doesn’t have as many supporters as its predecessor.

The iPhone SE2 wasn’t exactly a failure, it was just released at the wrong time – one didn’t catch the big screen trend, and the other didn’t support 5G networks. But Apple seems to have learned only half of its lesson by launching the third-generation iPhone SE in 2022.

The third-generation iPhone SE still opts for the same styling as the iPhone 8, with an upgraded processor to the A15 Bionic, and 5G network support as well, thanks to an enhanced processor NPU performance, and a Portrait Mode on the single camera. Although the addition of 5G network support, but the ‘old-fashioned’ styling, or let the iPhone SE3 is not marketed. Data research organization Canalys public information, iPhone SE3 total sales of only 24 million units, compared with the previous two generations of SE models in the same period of sales decline situation is obvious.

Despite the decline in sales, it can also be seen that the iPhone SE series still has a group of loyal consumers, this group of people may not care about fashionable styling and design, and no higher requirements for the image. In addition, long battery life, FaceID, MagSafe, are not in their scope of consideration, for this group of people, small, portable, performance is still acceptable, the “soul” home button is still there, it is enough.

For mass consumers, iPhone SE is more like a product designed by Apple for ‘inventory clearance’, no need to redesign the mold, directly replace the chip with a new one, the manufacturing cost is extremely low. However, judging from some of the changes above the iPhone SE3, it seems that Apple is trying to make it a cost-effective iPhone for occupying the mid-range market, but the reality is not as good as it should be.

Finally entered the full-screen era of the iPhone SE4, is likely to lose some of the consumers who once favored the SE series models, but like the design, performance and image part of the improvement, and the price compared to the base model of the iPhone is cheaper, it looks like it will be a good mid-range phone. But that leaves the SE lineup with even more challenges.

From the perspective of Apple’s smartphone lineup, the iPhone SE is still primarily a model that comes to fill the empty space in the $3,000 segment of the market, but with the sales of the base model iPhone getting more and more sluggish, Apple has ‘no choice’ but to give the SE a major overhaul.
The iPhone SE4 is indeed a mid-range model that fits the market environment, with decent performance, the same A-series chip as the base iPhone, and more importantly, a low price tag. But whether it will be recognized by mass consumers is still a big question mark to be drawn.

First of all, the iPhone SE4 uses the same mold as the iPhone 14, but the positive generation of the iPhone has entered the era of the “Spirit Island”, and with the entire iPhone 15 series “on the island”, there will inevitably be more developers to do new features and controls around this new design. Combined with the price, release time and performance performance, even consumers with a low budget can fully choose the base model iPhone 15.

Secondly, the imaging system of the new iPhone SE is still weak and not at all comparable to mainstream solutions. Imagine a new phone on sale in 2025, still using a single camera, even if Apple’s computational photography is strong, it is difficult to create a ‘miracle’ in the case of insufficient hardware level.

Last but not least, Touch ID is a major feature of the iPhone SE, and once the iPhone SE chooses to use the bangs screen design, most consumers will still focus on the base model iPhone.

For market considerations, Apple is bound to stabilize the iPhone SE in the mid-range market, and making so many changes in the fourth generation SE is also a careful preparation for taking a broader market. As for whether the iPhone SE4 can be accepted by mass consumers, we still have to see how Apple’s final pricing.

Translated with DeepL

World’s first satellite-enabled 5G call successfully dialed in Europe

On September 24, the world’s first satellite-enabled 5G call was successfully dialed in Europe by AST SpaceMobile’s satellite, as well as telecom operator Vodafone (Vodafone) and others. In addition, the test achieved a data transfer rate of 14 Mbps.

The 5G call was realized using the BlueWalker 3 test satellite, built by Vodafone, AT&T, and Nokia, which used a chip designed by British company EnSilica.

The 5G call, made from Maui, Hawaii, to a Vodafone engineer in Madrid, Spain, used spectrum provided by AT&T via the BlueWalker 3 satellite, and was made from an original Samsung Galaxy S22. The test broke the previous record for data speeds in space-based cellular mobile networks, achieving a 14Mbps The download rate was 14 Mbps.

Abel Avellan, CEO of AST SpaceMobile, said, “This achievement represents another major technological advancement and signals a shift in the way information is accessed.” Since the launch of the BlueWalker 3 satellite, the company has successfully achieved compatibility with all major cell phone manufacturers and support for 2G, 4G LTE, and 5G networks.

Tommi Uitto, President of Nokia Mobile Networks, congratulated the company, stating that space-based cellular mobile networks can provide vital network connectivity around the world, and he is proud of the key role that Nokia’s AirScale portfolio is playing in this network.

A new type of clock has a margin of error of one second in 300 billion years.

In the European XFEL X-ray laser, researchers have created a more precise pulse generator based on the element scandium, with an accuracy of one second over 300 billion years, which is roughly a thousand times better than that of the current caesium-based standard atomic clock. The team presented its results on September 27 in the journal Nature.

Atomic clocks are currently the most accurate timepieces in the world. These clocks use electrons in the atomic layers of chemical elements such as caesium as pulse generators to define time. Using microwaves of known frequency these electrons can be boosted to higher energy levels. In the process, they absorb microwave radiation.

Atomic clocks emit microwaves to the cesium atoms and modulate the frequency of the radiation to maximize the absorption of the microwaves; experts call this resonance. The quartz oscillator that generates the microwaves can be stabilized with the help of the resonance, allowing the cesium clock to be accurate to within one second over 300 million years.

Crucial to the accuracy of atomic clocks is the width of the resonance used. Current cesium clocks already use a very narrow resonance; strontium clocks are even more accurate, to within a second of 15 billion years. Further improvements are practically impossible to achieve using this method of electronic excitation. Teams around the world have therefore been working for many years on the concept of a “nuclear” clock, which uses the leaps in the nucleus of an atom as a pulse generator, rather than the leaps in the atomic shell. Nuclear resonances are much more violent than the resonances of electrons in the shell layers of atoms, but they are also much harder to excite.

At XFEL Europe, the team can now excite promising transitions in the nucleus of the element scandium, which is readily available as a high-purity metal foil or the compound scandium dioxide. This resonance requires X-rays with an energy of 12.4 kiloelectronvolts (keV, about 10,000 times the energy of visible light) and a width of just 1.4 femto-electronvolts (feV). This is 1.4 trillionths of an electron volt, which is about one-tenth of the excitation energy (10-19). This makes an accuracy of 1:10,000,000,000,000,000 possible.

“This corresponds to an error of one second in 300 billion years for this clock,” says Ralf Röhlsberger, a researcher at DESY working at the Helmholtz Institute in Jena, a joint institution of the GSI Helmholtz Center for Heavy Ion Research (HCIHI), the Helmholtz Zentrum für Dresden-Rossendorf (HZDR) and the Helmholtz Center for Heavy Ion Research (HZDR). Center.

There are many applications for atomic clocks that benefit from increased precision, such as precise positioning using satellite navigation. “The scientific potential of scandium resonance was discovered more than 30 years ago,” reports Yuri Shvyd’ko of the Argonne National Laboratory in the USA, the project leader of the experiment. “However, to date, no X-ray source has been able to emit light bright enough within the 1.4 feV narrow line of scandium,” says Anders Madsen, chief scientist at the European XFEL MID Experiment Station, where the experiment was carried out. Only X-ray lasers such as the European XFEL have changed this.”

In this pioneering experiment, the team irradiated a 0.025-millimeter-thick scandium foil with an X-ray laser and was able to detect a characteristic afterglow from the excited nucleus, which is clear evidence of scandium’s extremely narrow resonance line.

Equally important for the construction of atomic clocks is an accurate knowledge of the resonance energy, in other words the energy of the X-ray laser radiation at which the resonance occurs. Advanced extreme noise suppression and high-resolution crystal optics enabled the experimental scandium resonance energy value at 12.38959 keV to be determined to within five decimal places, a factor of 250 more accurate than before.

Jörg Evers, head of data analysis at the Max Planck Institute for Nuclear Physics in Heidelberg, emphasized: “The precise determination of the jump energy marks a major advance. An accurate knowledge of this energy is crucial for realizing a scandium-based atomic clock.”

The researchers are now exploring further steps towards the realization of such a nuclear clock. shvyd’ko explains, “The breakthrough in scandium resonance excitation and the precise measurement of its energy opens new avenues not only for nuclear clocks, but also for ultra-high-precision spectroscopy and the precise measurement of fundamental physical effects.

ChatGPT finally allows web searches, no longer limited to content by September 2021

OpenAI宣布，他们的聊天机器人ChatGPT现在可以使用微软的必应进行网上搜索，这使得ChatGPT能与微软必应和谷歌Bard平起平坐。

目前，OpenAI正努力争夺蓬勃发展的人工智能行业的主导地位。微软和谷歌已经让旗下聊天机器人进行网上搜索，并提供原始材料的链接，以提高信息共享的可信度和范围。但是，ChatGPT迄今为止只接受了有时间限制的训练数据，这些数据仅限于从互联网上收集的2021年9月之前的信息。在周三的一系列推文中，OpenAI表示，ChatGPT现在可以利用更多的最新信息。

OpenAI首席执行官萨姆·奥尔特曼（Sam Altman）在一条推文中说：“我们回来了！”在过去的几周里，OpenAI、谷歌、微软、亚马逊和Meta都在竞相推出生成式人工智能产品的更新和新功能。尽管许多新工具仍然会出现错误，但这表明这些公司正面临巨大压力，他们需要让用户使用这些工具生成新数据并提高性能。在OpenAI发布最新消息的同一天，Meta也发布了一款新的聊天机器人，同样可以使用必应进行网上搜索。

在此之前，ChatGPT付费用户可以通过链接其他公司开发的浏览器插件，让其进行网上搜索。周三更新使ChatGPT的核心产品能够直接进行搜索，这可能会扩大其用户范围。一些用户也可以在OpenAI运行的特殊测试程序中使用这个网上搜索工具。

微软是第一家推出联网聊天机器人的公司，今年2月推出了必应人工智能工具。由于与OpenAI签署了价值数十亿美元的协议，必应采用了OpenAI的技术。今年3月，谷歌推出了自己的聊天机器人Bard，并在部分用户的搜索结果顶部添加了聊天机器人功能。

OpenAI一直在快速开发面向消费者的工具，本周早些时候对ChatGPT进行了升级，使其能够生成图像并以语音回应，而不仅仅是文本。

Shots fired at Siam Paragon shopping center in Bangkok, Thailand

At around 16:00 local time on October 3, a shooting incident occurred at Siam Paragon Shopping Center in Bangkok, Thailand, with a number of Chinese tourists at the scene.
According to Xiao Zhou, a Chinese tourist who witnessed the whole incident, she was playing on the M level of the mall when suddenly three gunshots rang out from the toilet behind her. After the gunshots, people in the mall fled in all directions, panic spread to other floors and screams filled the mall. The young man, who was experiencing panic for the first time, described being “wrapped up in the crowd and running for his life.”
Local Thai media reported that the suspect in the case was arrested on the spot. Thai police revealed that the suspect, a 14-year-old boy with a gun, is currently under interrogation.
The shooting took place in the afternoon, when a thunderstorm was raging in downtown Bangkok, and a big traffic jam occurred in the city after the crowd poured out. Some Chinese tourists who could not get a taxi were sent back to their hotels by strangers, while others were not evacuated back to their accommodation until the early hours of the morning.
On October 4, according to the Chinese Embassy in Thailand, one Chinese citizen was killed and another injured in the shooting incident. At present, the victims have been identified, and the injured have been sent to the hospital in stable condition.
The Chinese Embassy in Thailand sent consular officials to the scene of the case at the first time to understand the situation, and went to the hospital to visit the injured, Thai Prime Minister Saita also went to the hospital to visit the injured Chinese.

Xiao Zhou and three companions have been traveling to Thailand for seven days, they plan to return home on October 4th. Their last day in Thailand was spent shopping at a mall in downtown Bangkok.
They chose the Siam Paragon Shopping Center, which is conveniently located near Bangkok’s BTS Siam station. Siam Paragon Shopping Center, which opened in 2005, was once the largest department store in Southeast Asia.
Thai time 16:00, she bought a bag in the shopping mall M layer, and friends, go to the bathroom next to the counter. 10 minutes later, Zhou just out of the bathroom soon, waiting for friends at the door, suddenly heard “bang, bang, bang” three consecutive gunshots. Xiao Zhou recalled that the gunfire from behind her inside the toilet, but also with echoes.
She was blindfolded in place and saw people in the mall screaming and running away, so she followed them. After running to the parking lot on level M, she heard another gunshot. She said she was afraid to stay in the parking lot and turned to run back, and saw a man in a restroom not far away beckoning her to hide in a bathroom stall.
Cho immediately ducked into one of the empty stalls and stood on the toilet, not daring to make a sound. She noticed that the cubicles next to her were filled with people from different countries.
Not long after entering the restroom, Xiao Zhou heard two more gunshots coming from the doorway. According to Xiao Zhou, this restroom was more than 100 meters away from the location where she first heard the gunshots, and this time the gunshots were louder and closer than the previous two, “My heart was about to jump out of my chest for fear of someone rushing in.”
After twenty minutes or so, another gunshot came from the other direction.
A few minutes later, there was no sound from outside. Gradually, someone went out of the cubicle, and Xiao Zhou stuck her head out to look outside and saw a uniformed Thai police officer, beckoning her to hurry. After Zhou went out of the restroom, she saw the crowd running without direction, and she ran all the way to the parking lot on the M level.
At the time of the shooting, Fu Qi (a pseudonym) was eating dessert with her boyfriend on the first floor. Suddenly, she heard a “loud bang”, which she first thought was a heavy object falling, then another bang, and the crowd began to run wildly.
Looking at the situation, she hesitated “for a second between hiding under the table and running away,” before being pulled by her boyfriend and running wildly outside the mall. She saw someone fall on the road, but had no time to care.
On the second floor of the mall, Ms. Zhang and her husband were on their honeymoon in Thailand, and had just gone up the escalator to go downstairs to buy gifts for their family when she saw someone running on the first floor. Ms. Zhang remembers that her husband yanked her while shouting “don’t go down, come back”, followed by two or three gunshots that she heard.
Ms. Zhang was so scared that she started to turn around and ran up the escalator, and because she was too panicked, she fell on the elevator and her leg hit the corner of the elevator. She was helped up by her husband several times before getting up.
As she recalled, when they ran back, there were quite a few people on the second floor who still didn’t know what had happened.
Ms. Zhang’s husband pulled her toward the second floor parking lot and ran, shouting “run” as they ran. “But there was a language barrier and everyone was in a state of confusion.” Then the crowd began to disperse, surrounded by screams.
Running to the intersection with the parking lot, Ms. Zhang’s knee began to bleed, her husband stopped to wipe her blood, when the crowd began to pour over, the two of them continued to run.
Fu Qi said she did not see the murderer, but was “wrapped up in running for her life”, and after fleeing the scene, the fear gradually came.
Fu Qi and her boyfriend ran outside the mall after heavy rain began to fall in Bangkok, and she said she saw people constantly running out and police cars and ambulances coming.
Min Chan and her family made their way along the mall’s security corridor, past the parking lot, and eventually escaped. Outside the mall, people were still screaming and running away. Chen Min and her family ran to another nearby mall before they dared to stop.
It was raining heavily outside, the road was waterlogged and the whole road was blocked. Zhou, who escaped, pulled open the door of a car and jumped in. Luckily, she met two local Thai girls who understood what Zhou was going through. Because she did not speak Thai, Xiao Zhou used cell phone software to communicate with them, and after more than an hour of traffic jam, she was sent to the hotel by the two girls.
Ms. Zhang and her husband both ran to the second floor parking lot and ran down to the road from the second floor parking lot.
She saw that there were still people outside who were ready to enter the mall, so she and her husband both ran and shouted in Chinese and English respectively, “Don’t enter, run.” Ms. Zhang recalls, “But again, the language wasn’t clear, and not everyone could understand us.”


The rainstorm got heavier and heavier, and Ms. Zhang and her husband didn’t get a taxi for a long time; later, they stopped a tricycle to leave. From the tricycle, they saw a police car heading in the direction of the incident.
She remembers that when they stopped to wipe the blood from their knees in the second-floor parking lot, crowds of people rushed by. There was a foreign girl who tugged on her arm as she passed by, saying in English, “Hurry up,” which she remembers still feeling touched, “I’m very grateful to her.”
On October 4, local time, the head of the Kyoto Police Department in Bangkok, Thailand, said police filed five preliminary charges against the suspect (male, 14 years old) who created the Siam Paragon shooting case, namely, intentional homicide, committing homicide, unlawful possession of a firearm, carrying a firearm into a public place without permission, and firing a firearm in a public place without permission.
Police said they are gathering more evidence on the other charges, and are currently investigating whether the suspect’s guardian violated the Child Protection Act, and prosecutors, medical and psychological experts are investigating to assess the suspect’s psychological state. The suspect is being held at the detention center of Bangkok’s Pathumwan police station and will be escorted to the juvenile court for trial on the afternoon of the 4th.


In addition, police have searched for bullets in the suspect’s home, which will be included in a separate investigation.
Chinese citizens injured in Bangkok, Thailand shopping center shooting in stable condition
According to CCTV News, on October 4, local time, the director of the Emergency Medical Service Center in Bangkok, Thailand, Yutana announced the treatment of the injured in the shooting incident at the Siam Paragon shopping center. Yuktana said that after the incident, the center transferred a total of 7 people, 1 person died on the spot, 2 people were sent to the police hospital, 1 person died on the way, 3 people were sent to Chulalongkorn Hospital, 1 person was sent to Bangkok Christian Hospital. Two of the five injured are still in the intensive care unit of the hospital. The condition of the one injured Chinese citizen is stable.

NVDIA RTX 50 ,Return of the King!

According to the roadmap, NVIDIA will launch the next-generation Blackwell GPU architecture in 2025, and blossomed into two: GB100 series cores for high-performance computing and artificial intelligence, and GB200 series cores for gaming and creation.

According to the latest news from authoritative exposure expert kopite7kimi, the GB100 big core will be equipped with 10 groups of GPCs, and each group of GPCs is further divided into 8 groups of TPCs. this is not the same as the layout of GA100 8 x 8, GH100 8 x 9.

Theoretically, each TPC group should still have 2 groups of SMs, and each SM group should still have 128 CUDA cores, that’s a total of 20,480 CUDA cores, which is an 11% increase over the current GH100 cores.

At the same time, the bit width of the HBM memory has been expanded to 8192-bit, which is a full 1/3 increase compared to the GH100, meaning that the memory capacity can be up to 128GB, and will inevitably be upgraded to HBM3.

Of course, as usual, B100 cards will shield some units and memory bit width/capacity.

We’re more interested in the gaming card core GB202, which is said to have the same 12 GPC groups as the current AD102, but with 8 TPCs per group instead of 6.

This means that GB202 will have a total of 192 groups of SMs and 24,567 CUDA cores, a full 1/3 increase!

Even more amazingly, kopite7kimi confirms previous speculation that the GB202 core will really support 512-bit bit-width, again an increase of 1/3 from the current level, which means that the video memory capacity will be up to 32GB, and it will be the new generation of GDDR7.

The last time there was a native 512-bit memory bit-width, it was the AMD R200/R300 era Hawaii GPU cores. As usual, the RTX 5090 won’t be full-blooded, nor will there necessarily be an RTX 5090 Ti.

According to previous claims, the GB202 core will also have a doubled 128MB cache and a core frequency of up to ~2.9GHz, which can actually exceed 3GHz. It will also have four different levels of cores underneath it such as GB203, GB205, GB206, and GB207.

South Korea’s military parade unveils unmanned weapons on land, sea and air in a big way: the drone matchup in East Asia

On September 26, the South Korean army held a military parade to commemorate the 75th anniversary of the founding of the army, displaying a large number of weapons and equipment, in addition to those conventional weapons often exported by South Korea, but also displayed a number of drones and unmanned aircraft weapons, the army, land, sea and air.

First of all, explain the anniversary of the founding of the South Korean army, South Koreans call it the Korean National Army Day, yes, but also the National Army, National Army Day dates back to August 15, 1948, the Americans will be the southern part of the peninsula of the Guard and Coast Guard integrated into the Korean National Army.

But National Army Day is set for Oct. 1 because that’s the day the U.S. and ROK troops crossed the 38th parallel and attacked to the north.

However, the parade is often held to avoid the Mid-Autumn Festival holiday, so this year it was held on September 26th.

What was notable about the ROK’s current parade were the drone weapons used by the Army and Navy.

For example, the Army was accompanied by an unmanned vehicle in the Korean version of the Hummer.

This kind of unmanned vehicle for South Korea’s own development, called ARION-SMET, for 6×6 unmanned vehicles, but also participated in the U.S. Army unmanned vehicle bidding test, and do not know the U.S. Army will finally buy.

ARION-SMET is the English acronym, the full name is “cross-country intelligent navigation of the automatic robotic system – small multi-functional equipment transportation”.

And ARION is the name of a steed in ancient Greek mythology.

This naming system is very similar to that of the U.S. military, and it seems to be aimed at the U.S. military procurement.

The ARION-SMET is based on a 4×4 civilian unmanned vehicle, which is purportedly fitted with a remotely operated weapon station using deep neural network technology that can detect enemy targets and attack them, moving with the infantry company.

In addition to fire support, it can also transport weapons and ammunition, reconnaissance and transport the wounded.

Its maneuverability is a maximum speed of 43 kilometers per hour on the road and 34 kilometers per hour off-road, with a maximum travel of 100 metric centimeters and a maximum load of 550 kilograms.

The ROK’s unmanned combat ships were paraded on trailers, with a total of two types of surface unmanned combat boats and one unmanned submarine.

The surface ship on the left is the “Sea Sword-II” (해금-II) unmanned combat ship, first made public by South Korea in 2019, 12 meters long, 3.5 wide, fully loaded with 11 tons, 2 diesel engines, a maximum speed of 35 knots, 20 knots of the maximum range of 180 beavers.

There is a 12.7mm machine gun remote controlled weapon station, and 70mm multiple rocket launcher at the stern, mounted radar and other detection equipment.

Mainly used for patrols near military bases, there are also armed reconnaissance missions to the sea radar blind zone.

The right surface ship is “water ghost” (네이비 씨 고스트), public information is not much, from the South Korean Navy in the South Korean Navy base in Zhenhai on November 11, 2022 in the 77th anniversary of the founding of the South Korean army ceremony related information can be seen, this is a kind of unmanned boat can be mobilized, but also can be manned by manned ships.

From the parade picture speculation, it should be similar in size to the “Sea Sword – II”.

The name of the South Korean army’s unmanned submarine is very simple and crude, called ASWUUV, is the English acronym for anti-submarine underwater unmanned combat vehicle, in June 2022 had an operational test.

The ASWUUV is 6.5 meters long, has a maximum displacement of 9 tons, a maximum cruising speed of 10 knots, a maximum dive depth of 300 meters, an endurance of 30 days, and is equipped with passive and active sonar, acoustic sensors, a communications unit and a hydrogen fuel cell.

South Korea had unveiled the concept of these unmanned combat vessels in defense exhibitions, which is to be carried to the operational waters by a mothership and then formed into combat clusters to carry out combat missions.

On the drone front, a number of models were also on display.
The first is the KUS-FS/KUS-15, a medium altitude UAV that resembles the US RQ-4 Global Hawk.

It is said to be a UAV that South Korea started developing in 2011, and some people say that this UAV is called a MUAV, which is not true!

MUAV is the English acronym for Medium Altitude Unmanned Aerial Vehicle.

It has a wingspan of 25 meters, a length of 13.1 meters, a maximum flight time of about 24 hours and a maximum altitude of 11.9 kilometers, and is mainly used for reconnaissance, with the main reconnaissance equipment being a synthetic caliber radar with an antenna of 1.2 meters in diameter, and is currently assigned to the 39th Reconnaissance Wing, which will be established in 2020.

The South Korean side has already begun to add weapons to the KUS-FS test, installed a weapon mount on each side of the wing, has successfully conducted a conventional bomb drop test, the future may be added to the anti-tank missile launch capability. In other words, it may be developed into an investigation and combat UAV. The second one looks like a U.S. MQ-9 Predator drone.

According to public information, this drone is actually still under development, and during the military parade, this drone had obvious shaking in front of the camera, most likely a model.

This drone is tentatively named the next generation UAV (차기 군단급 무인기-II or UAV-II), and is designed to have a cruising radius of 600 kilometers, a maximum takeoff weight of 1,700 kilograms, a maximum flight time of 18 hours, and is equipped with a synthetic aperture radar with satellite communication capabilities.

One of the unmanned helicopters is a previously undisclosed model that differs greatly in size and shape from the already disclosed KUS-VH unmanned helicopter, so I wonder if it’s a downsized model or a real drone?

Behind the unmanned helicopter are some small drones, the aerodynamic shape resembles South Korea’s first self-developed drone, the RQ-101/Night Invader-300, but the size doesn’t match and the writing on it should be FT-3.

So this would be the FT-3 of South Korea’s KUS-FT series of drones, there isn’t much publicized detailed performance data, but the public performance of the FT-2 from this series is:
3.7 meters long, 4.5 meters wide, 0.9 meters high, maximum takeoff weight of 150 kg, the engine is 35 hp, maximum speed of 200 km / h, cruise speed of 130 km / h.
It is mainly used for reconnaissance of ROK Army divisional units.

Behind the FT-3 is the KUS-7 UAV, which is used for reconnaissance and artillery unit calibration. It has a length of 3.4 meters, a wingspan of 4.2 meters, a cruising speed of 150 km/h and a flight radius of 50 km.

There are also a number of stealth-designed drones in the drone parade.


In terms of appearance, it looks like the KUS-FC unmanned fighter aircraft previously displayed by South Korea, but according to South Korea’s own public information, the KUS-FC unmanned fighter aircraft has a wingspan of 10 meters alone, which clearly does not match this.

So it is possible that these are models of drone fighters?
At the end of the drone parade are Harpy drones, Israel has exported Harpy drones to South Korea before, I wonder if these are imported Harpies or South Korean copies?

Guo Mingji: Next year, the shipment volume of EUV lithography machines will plummet by 30%!

Recently Ming-Chi Kuo, an analyst at TFISEC(TF International Securities Group) , said his latest research indicates that lithography maker ASML may sharply cut its 2024 EUV equipment shipment forecast by about 20-30 percent.

According to Kuo, demand for advanced chips and the devices they are needed for may be lower than expected in 2024 due to weak sales of MacBook laptops and iPads. He also predicted that memory chip makers won’t go ahead with planned factory expansions until 2025 at the earliest, which will lead to weaker demand for ASML lithography.

On the question of when the chip industry will recover, Kuo wasn’t sure, saying it could take longer than expected. “The current market consensus is that the semiconductor industry will bottom out in the second half of 2023. However, close attention needs to be paid to whether this bottoming out will be delayed to 1H24/2Q24 2024.”

In April of this year, there were rumors that ASML’s photolithography had also suffered a significant order cut as a result of TSMC’s scaling back of capital expenditure, and that its year-on-year orders for 2024 threatened to be slashed by more than 40%. Although the report did not specify whether TSMC’s purchase of ASML’s 2024 equipment was reduced by 40% or whether ASML’s overall 2024 orders were slashed by 40%, the overall report seems to refer to the former.

Other reports say that it’s TSMC’s 2024 EUV lithography purchases for ASML that have been cut (or delayed) by 40%.

It’s worth noting that during the Q2 FY2023 earnings meeting in mid-July this year, ASML CEO Peter Wennink said, “As for the EUV system, we’re seeing some shifts in demand in terms of timing. It’s driven in large part by factory readiness. The manufacturing plants are not ready yet, so we will ship later (here seems to confirm previous rumors that TSMC asked for a delay in EUV shipments). This means that the initial 40% growth in EUV revenues will drop to about 25% this year.”

In mid-September, Reuters, citing two people familiar with the matter, reported that TSMC had notified key suppliers to delay shipments of high-end chip manufacturing equipment, including EUV lithography, due to growing uneasiness over sluggish customer demand.

Days before the release of the report, Peter Wennink in an interview with Reuters also revealed that some orders for ASML’s high-end equipment has been delayed, but he did not point out the name of the customer, which he said would be a “short-term management” problem, “We have seen several reports about the readiness of the fab. We’ve seen several reports on fab readiness, not just in Arizona…but also in Taiwan,” he said.

Goldman Sachs Securities estimates that TSMC 2023 capital expenditures remain at about $ 31.6 billion, and no adjustment, but because of the research and judgment demand recovery is not clear, TSMC is likely to slow down to advanced process demand-based equipment procurement pace, instead of part of Taiwan’s equipment to Japan, the United States, and other overseas production bases to use, so Goldman Sachs TSMC 2024 capital expenditures are expected to drop from $ 28 billion to $ 25 billion, which is equivalent to the same as the TSMC 2024 capital expenditures from $ 28 billion to $ 25 billion, which is equivalent to the same as the TSMC 2024 capital expenditures. Therefore, Goldman Sachs on TSMC 2024 capital expenditure is expected by 28 billion U.S. dollars to 25 billion U.S. dollars, in the form of 2023 than a 21% reduction; as for the capital expenditure in 2025 from 36 billion U.S. dollars to 35 billion U.S. dollars.

In addition, Goldman Sachs Securities also synchronized downward revision of TSMC 3 nm process capacity utilization, 2023, 2024, respectively, from 40%, 71%, down to 36%, 65%, 2025 is expected to maintain the 78% unchanged. 3 nm process in 2024, 2025 capacity, also from 80,000, 90,000 tablets per month, adjusted to 70,000, 80,000 tablets per month.

However, in July of this year, the second quarter earnings meeting, Peter Wennink has said that as of the time of the earnings release, ASML has a backlog of about 38 billion euros, which provides a good basis for dealing with short-term uncertainty. Despite the increased uncertainty, ASML still expects strong growth in 2023 due to the strong growth trend in the DUV lithography business. Compared to 2022, revenues are expected to grow by 30% year-on-year in 2023 and gross margins are expected to improve slightly.

During the time period when the old and new Tesla Model 3s are alternating, there are a lot of wait-and-see consumers who put down deposits but are still torn between choosing the old model or the new one.

When the price of the new model came out, the users who might have chosen the old model smiled, a decision that saved them 28,000 yuan.Then K brother to choose the new model of the enthusiasts to play a strong shot, the new model is worth the extra 28,000 yuan you wait. Or let me put it more directly, this wave is a win for the wait party!


The reason for this is that we’ll start with the exterior and interior, as usual. This time, K got one of the first test drives of the new Model 3, so the experience will be a bit richer.

The front end is the part of the new Model 3 that has changed the most, the lines have become flat and sharp, the headlights are more narrow and long, and the daytime running lights look the best when they are on..


The overall profile of the body from the side is largely unchanged, but the suspension is noticeably taller than the old model, and it’s surprisingly possible to fit a fist between the tires and the wheel arches!If the transport mode hadn’t been unlocked, then I’d suggest

the shocks could have looked ahead.

The rear gets a pair of new, more integrated taillights, with a more continuous “C” light source and a more subtle look. The rear badge is now “Model 3+”, which is exclusive to the longer-range version.The diffuser at the bottom is larger, and with the changes to the front, the wind resistance of the new Model 3 has been lowered by 0.01 cd, and it’s said that every 0.01 cd reduction in wind resistance increases the range by 10km..


The changes to the exterior aren’t much, the “optimization” inside is a bit more, why the quotation marks?


Just take a look at the steering wheel of the new Model 3, and you’ll see that it doesn’t have any shortcut multimedia buttons, but rather turn signals, low beams, and windshield wipers!


K tried it out on the lot, and it worked, but didn’t like it..
My hands are always unconsciously wrenching towards the air behind the steering wheel, including shifting gears, so imagine for yourselves.


That’s right, the gearshift lever is gone too, in favor of sliding the gearshift on the left side of the screen, slide up for forward, slide down for reverse. It’s a good thing the physical buttons remain on the roof for safety, and you can shift gears by touch.Pictures
emm, this wave of operation is very difficult to comment. Other manufacturers are discouraged from imitating it.If it weren’t for the eye-catching 256-color ambient lighting, I’d be staring at those two disappearing levers.


Ventilation has been added to the seats across the board, and it’s one of the changes that’s perceptible as soon as you get in the car.


On hot summer days, the ventilated seats must work, and the seats are so comfortable that the new Model 3 is really the only one that feels at home in a rough house!Pictures
The audio is differentiated between high and low end, with only 9 speakers in the standard version and 17 in the long range version.As far as the experience of the long range version is concerned, the sound is really much better, the bass is strong enough and the vocals are very clear. Get it? It’s explicitly and implicitly pushing you to buy the Model 3+!The back row gets an epic update with the surprising addition of a screen!.


I wonder if a screen is an inch of gold in the eyes of an old horse?But from the official explanation, the main purpose of the screen is to give the rear passengers control over things like airflow, music, seats, etc., with the entertainment features being secondary.The rear space has not changed, it is still quite compact, family users choose Y a little more suitable.PicturesThese are the static parts of the real-world experience, next take you on the highway, where you can’t see, the chassis is the most valuable upgrade of the new Model 3!


If you’ve driven the old Model 3, even the later version with blacked-out window frames, the suspension is still straight and flinty, with four suspension bars and chopsticks, making direct contact with the road, and the sense of luxury will be a little worse.But the chassis of the new Model 3 is not to be missed, and a thin layer of suspension has been added!


Even on “flat” roads, you can feel the suspension working hard, filtering out uncomfortable vibrations with a slight rocking motion, and the chassis is quite heavy and doesn’t feel floaty.
You could say that the new Model 3 is a luxury sedan, without the cheapness of a “gadget”.


If you think comfort and sport are mutually exclusive, you’ve never driven a 3 Series with an F-chassis.


The new Model 3 is the pinnacle of balance, with a suspension that remains supportive during dynamic cornering, and steering that’s very agile and unclunky.
It’s a shame that the tires have been downgraded to Hantaes, as the new Model 3 would have more limits if they were Michelin PS series.Pictures.In terms of power, the long range version is equipped with dual front and rear electric motors, with a total electric motor power of 331kW, a total torque of 559N-m, a zero to one hundred acceleration time of 4.4 seconds, and a combined CLTC range of 713 kilometers.


Really, take my advice, it’s more than enough.The standard range is the same as the old model, the motor power is exactly the same, the zero to one hundred acceleration time is 6.1 seconds, and the CLTC comprehensive range has been improved by 50 kilometers to 606 kilometers.Intelligent driver assistance this piece of change is also quite big, the whole car’s radar are eliminated, using pure visual camera program. Even if you don’t use the FSD, reversing is only supported by s.


At present, the reversing only has distance alerts, no sound warning, this point still needs to be improved.The vision-based FSD program can’t be tried yet, so K will be the first to experience it when he has the chance.After all is said and done, the new Model 3 is still the most worthwhile to take out the chassis part of the changes, and the old model comparison is two completely different cars, if the old model is a palm mattress, then the new model will be upgraded to Simmons.To use an old-fashioned word, the new Model 3 is a lot more “advanced” to drive.


This kind of unseen change is precisely the new car’s biggest killer feature, and it will make it easier to deal with strong competitors like the ET5.For consumers, the situation is now clear.Those who want a taste will just order the long range version, which is by far the most flavorful one.If you still want to eat the price cut dividend like the old Model 3, you can also wait a bit, the price cut is inevitable, just no one knows when this will happen.

OpenAI wants employees to sell shares to raise money.

OpenAI is reportedly in discussions with investors about a possible stock sale, with the artificial intelligence startup seeking to sell shares at a valuation of $80 billion-$90 billion (roughly Rs. 584.248 billion – Rs. 657.279 billion at current exchange rates), a valuation that nearly triples the amount it was valued at earlier this year.

The deal is expected to allow employees to sell their holdings rather than raise additional capital in the form of new shares issued by the company, the WSJ said. OpenAI has already begun convincing investors, people familiar with the matter said, adding that OpenAI revenue is expected to reach $1 billion this year and billions more by 2024.

The $1 billion figure is consistent with the numbers that broke in the media in August. OpenAI reportedly generates $80 million in monthly revenue, up from $28 million in all of 2022. ChatGPT Plus, its $20-per-month paid version of ChatGPT that launched in February, has fueled OpenAI’s revenue growth.

On the other hand, a stock sale would allow employees to understand the value of their equity without having to wait for the company to go public, could help the company attract top talent and generate liquidity, and would bring OpenAI a new valuation. an $80 billion or higher valuation would make OpenAI one of the most highly valued startups in the world, behind only ByteDance and Musk’s SpaceX.

More money raised, but more money spent
OpenAI aims to sell hundreds of millions of dollars worth of its existing stock to Silicon Valley investors, the report said. In April, OpenAI raised more than $300 million from investors including Sequoia Capital, Andreessen Horowitz, Thrive and K2 Global, valuing the company at $29 billion. This is unrelated to a large investment announced by Microsoft earlier this year, which was completed in January of this year, with an investment size of about $10 billion. By some counts, OpenAI had already received $4 billion in investments in its seven-plus years of existence before it received the $10 billion investment from Microsoft. At this point, OpenAI’s cumulative financing has amounted to 14.3 billion U.S. dollars (about 104.462 billion yuan at the current exchange rate).

While it was reported back in May that OpenAI was looking to raise more money, the share sale, if it continues, will not provide OpenAI with additional working capital, but will only allow its employees to divest some of their shares.

In May, The Information reported that three people with knowledge of OpenAI’s finances had revealed that OpenAI’s losses had roughly doubled to around $540 million as a result of the development of ChatGPT and the hiring of key staff from Google last year. Dylan Patel, principal analyst at consulting firm SemiAnalysis, estimates that it costs $700,000 per day to run ChatGPT.

According to Fortune’s disclosure, the company’s $544.5 million in total expenses for 2022 consists of $416.45 million in compute and data expenses, $89.31 million in employee expenses, and $38.75 million in other operating expenses for which no specific programs were identified. Those costs were racked up before the $10 billion investment Microsoft made earlier in the year.

OpenAI has not stopped upgrading and updating its products, and it’s clear that the huge cost investment continues:

In the early morning hours of September 21st, OpenAI announced that its text-based diagramming tool, DALL-E, will soon be upgraded to DALL-E 3, and will be natively integrated into ChatGPT. Compared to last year’s DALL-E 2, DALL-E 3 is significantly better at understanding text and generating images with the same prompts. The often criticized problem of “not being able to generate text on images” has also been solved in this update.

On September 26th, OpenAI launched multimodal ChatGPT, which, in addition to the usual text box interactions, will introduce voice and images to Plus and Enterprise users, allowing them to have a conversation with their voice or show ChatGPT what they are talking about. Voice will be available on iOS and Android (opt-in in your settings) and images will be available on all platforms.

Also on August 16, OpenAI announced that it had acquired the team at Global Illumination, its first foreign acquisition since its founding in 2015, but did not disclose the amount of money involved in the deal.

While the company’s revenue grew after OpenAI launched a paid version of its chatbot in February, those costs are likely to continue to rise as more customers use its AI technology and the company trains for future versions of the software. In fact, Microsoft and other recent investors have covered most of these costs, but they may stop investing if OpenAI doesn’t turn a profit soon.

Meanwhile, on August 24th, media reported that OpenAI CEO Sam Altman will travel to Abu Dhabi, the capital of the United Arab Emirates, and other places in the second half of the year to seek financing, which will be on a huge scale of no less than $100 billion.

It is reported that the story Altman told to VCs is not limited to AGI general artificial intelligence, he said that OpenAI’s goal is to realize Super intelligence (Super intelligence), for example, can attack cancer in a month. But OpenAI is still very far from that goal, and the amount of money OpenAI needs to achieve it is unimaginable.

For now, Altman has no plans to raise money through an IPO. In June, Altman said he wasn’t interested in going public because he didn’t want to be sued by the public markets, Wall Street, and so on. He explained that when developing AI, he may make some decisions that seem very strange to public market investors.

The biggest investor, too, can’t be pitched
Microsoft took a 49 percent stake in OpenAI after investing $10 billion in the company earlier this year. As part of the deal, OpenAI promised to work with it and integrate its AI software with Microsoft’s products in exchange for computing resources to train and run its models. Microsoft has indeed raced to build AI capabilities into most of its software products, including the GPT-4-based Windows Copilot, but today, big models are putting pressure on Microsoft as well.

The company is concerned that with Windows having more than a billion users worldwide, the cost of running these AI features could grow rapidly. And Microsoft doesn’t want to give up the financial benefits of its new AI offerings, so it’s looking for low-cost alternatives.

In recent weeks, Peter Lee, who oversees Microsoft’s 1,500 researchers, instructed many of them to develop conversational AI, which may not perform as well as GPT-4 but is smaller and much cheaper to run, according to people familiar with the matter.

Microsoft’s research team reportedly has no illusions about developing large AI models like GPT-4. The team doesn’t have the same computational resources as OpenAI, nor does it have a significant amount of human labor to provide feedback on the questions answered by LLM for engineers to improve.

Microsoft’s efforts to move to more efficient AI models are still in the early stages, though the company has reportedly begun testing internally developed models in services like Bing Chat.

Mikhail Parakhin, head of Microsoft’s search division, has previously said that Bing Chat relies 100 percent on GPT-4’s creativity and accuracy model. However, in its balanced model, it uses a new model called Prometheus and a Turing language model. The latter are not as powerful as GPT-4: they can recognize and answer simple questions, but when they are faced with tougher ones, they pass them on to GPT-4.

On the coding front, Microsoft recently unveiled its 1.3 billion-parameter Phi-1 model, which is said to be trained on “textbook-quality” data and generates code in a much more efficient way, but it’s not quite up to GPT-4 standards.

The company is also working on other AI models, such as Orca’s open-source Meta-based Llama-2 model. According to the report, Orca’s performance is close to that of OpenAI’s model, although it’s smaller and consumes fewer resources.

Microsoft’s AI research division has about 2,000 graphics cards from NVIDIA available for use, and Lee has now ordered that most of them be used to train more efficient models focused on performing specific tasks, rather than the more general-purpose GPT-4, according to the report.

There’s no denying that OpenAI and other developers, including Google and Anthropic, are ahead of Microsoft in developing advanced LLMs, but Microsoft may be able to compete in the race to build models of similar quality to OpenAI’s software at a fraction of the cost.

Translated with DeepL