The recently released artificial inteligence software ChatGPT iOS app is getting a new update. The new version offers practical new functions for iPhones and is now also available for iPad users.
Image: Artificial Intelligence, AI (Inteligencia artificial), via: pixabay, by: geralt.
The recently released iOS app version of the artificial inteligence software ChatGPT is getting a new update. This new version will offer some practical new functions for Apples iPhones and is now also available for Apple iPad users.
The official iOS app for ChatGPT gets a new interesting update. Because with the latest version, users can now also use the app on their iPad devices. They will also be able to share content via drag and drop. This makes using the AI-based chat platform much easier and offers new possibilities for interaction.
ChatGPT for iOS: Verbal voice input with Siri is now also possible
A special highlight of the update is the Siri and Shortcuts support. With a predefined shortcut, users can now set custom commands to launch the ChatGPT app by their voice. Also they can instantly share their questions or concerns. This makes using the app very practical and intuitive.
The hype surrounding artificial intelligence (AI) has increased significantly in recent years. AI is also increasingly finding its way into everyday work. According to a report by Goldman Sachs, AI will finally hit the mainstream in 2023 and is now “too big to ignore”.
Companies like OpenAI have gained great popularity with AI tools like the now world-renowned chatbot ChatGPT or the image generator Dall-E. Content generated by a bot quickly became the norm.
No wonder, then, that the demand for AI-enabled chips has risen sharply recently. Chip manufacturers like Nvidia benefit from this boom, because AI programs like ChatGPT require a lot of computing power, and GPUs are ideal for these requirements.
Nvidia at the top
With a share of 80 percent, Nvidia dominates the world market for graphics card chips and offers a wide range of products specifically for machine learning. “If the AI trend proves sustainable, the immediate demand will be for chips and computing power, and that’s where Nvidia is the flagship at the moment,” Thomas Hayes, chairman of private equity firm Great Hill Capital, said of Nvidia’s strong gains, according to german news Tagesschau the past few weeks.
But not only Nvidia benefits from the AI hype. For example, AMD, a direct competitor of Nvidia, specializes in developing microprocessors and chipsets. The company has announced that it will make its AI accelerator business its top priority, which could soon rival Nvidia.
Contract manufacturers such as the Taiwan Semiconductor Manufacturing Company (TSMC) and software companies such as Service Now also work closely with Nvidia and are benefiting from the AI boom.
The stock markets are not unaffected by the hype surrounding artificial intelligence. Logically, the most popular stock among professional investors is Nvidia, but other AI-related stocks are also enjoying increasing popularity. Companies like Broadcom, TSMC, and Service Now have all rallied in recent months, significantly increasing their profits. ASML and Marvell were also able to post clearly recognizable growth.
According to forecasts by the Green Software Foundation, information and communication technology will account for around 20 percent of all electricity consumption by 2030. Emissions from digital technologies will double by 2025 compared to 2019 levels.
But the technology industry is becoming increasingly aware of its carbon footprint. Last but not least, against the background of the energy crisis, the importance of green IT is becoming increasingly apparent.
Green IT summarizes all measures that combine technological progress with environmental protection. A distinction is made between Green by IT and Green in IT. Green by IT are technologies that actively help to achieve sustainability goals. Such as software that makes consumption measurable and shows potential for optimization. Green in IT, on the other hand, aims to optimize IT processes. So that they have the least possible negative or even positive impact on the environment and resources.
This is not primarily about limitations, but about responsible and resource-saving use of technology. The greatest possible benefit should be obtained from every gram of CO₂ emitted into the atmosphere. This enables the demand shaping principle in software development.
Demand shaping
Demand shaping is a strategy to influence demand to match existing supply. Accordingly, when supply is low, demand is reduced and increases with supply accordingly. An example of this is video conferencing. When the user has low bandwidth, the video quality is reduced while the essential audio quality remains high. Demand (video quality) is adjusted to match supply (bandwidth).
Another example of demand adaptation is progressive enhancement in web design. The most basic form of a website is made available for older browsers and with low bandwidth. The more resources and bandwidth a user has available on their device, the more features are provided. But these are optional.
This principle can also be used to achieve energy efficiency. The energy requirements of applications are matched to availability. Demand shaping is therefore opposed to the widespread over-provisioning principle of providing more resources than are necessary to cover peak loads or increasing demand.
Through demand shaping, so-called “eco modes” can be built into software applications. Similar to those in cars and household appliances. The application can be used in an emissions-friendly way at the expense of performance or at full power with higher energy consumption. Applications can either be set to eco mode by default, or users can choose. Based on the nudging principle.
Another example of sustainable applications are applications optimized for edge computing. Data and process steps or complete applications are brought closer to the users instead of being processed in remote data centers. This not only reduces latency, but also CO₂ emissions, since less energy is required to transmit the data.
Renewable energy
Applications can also be programmed in such a way that the respective mode – energy saving or maximum performance – is made dependent on the availability of renewable energies.
Demand shaping is thus related to the principle of demand shifting, i.e. the shifting of demand. Here the demand for computing, storage or network resources is shifted to other regions or to times when the availability of renewable energies is higher. Companies should rely on solutions that automatically move computing, storage and network resources to where the carbon footprint is lowest.
Both demand shaping and demand shifting are important to reduce CO₂ consumption in IT. Depending on the application, developers should determine whether the computing power of applications should be reduced or relocated if the CO₂ intensity is high.
After delays, a preliminary version of Apple Pay Later has now been launched for selected users in the USA. This should be able to split payments into four installments.
Apple Pay users can now also pay in installments – at least in the USA. There, the tech giant from Cupertino has now launched a pre-version of Apple Pay Later. This was announced by Apple via a company announcement .
You can now pay in installments with Apple Pay. (Photo: nikkimeel/Shutterstock)
Pre-release version for select US customers only
However, even in the United States, initially only randomly selected users can benefit. You will receive an invitation for the pre-release version. Customers who want to enjoy this experience in the US must also have an iPhone with the recently released iOS 16.4 or an iPad with iPadOS 16.4.
It is not yet clear when the full version will start in the USA. Apple is talking about the next few months.
Apple had already presented Pay Later in June 2022 at the Worldwide Developers Conference 2022. However, due to delays caused by alleged technical problems, the playout was pushed back. Then, earlier this year, Apple tested the feature in beta, first by employees and then by retail staff.
This is how installment payments via Apple Pay work
With Apple Pay Later, users can split payments into up to four installments. These must be paid within six weeks. Interest and fees do not apply. In theory, you can pay at all retailers that support Apple Pay.
The loans that can be applied for through Pay Later range from a minimum of $50 to a maximum of $1,000. According to Apple, a “gentle credit check” runs in the background for every transaction.
Refunds can only be processed via debit cards. Credit cards are not accepted as they could send customers deeper into a credit spiral.
Management of installments via Apple Wallet
Users can track and manage when the installments are due via Apple Wallet. Pay Later is fully integrated into the app. Just before the installments are due, Wallet sends a notification to the user.
To ensure security and privacy, Apple Pay Later authenticates transactions via Face ID, Touch ID, or passcode.
The powerful language model ChatGPT generates texts that can hardly be distinguished from those of human authors. We explain the technology behind the hype.
Since the US company OpenAI released its new artificial intelligence (AI) ChatGPT for free testing at the end of November last year, users on social media have been sharing masses of examples of how the chatbot answers knowledge questions, formulates e-mails, writes poems or texts summarizes.
ChatGPT’s ability to confidently deal with natural language and to understand complex relationships with a high hit rate is seen by some observers as another milestone on the way to strong artificial intelligence – i.e. to algorithms that are on a par with human thinking ability in every respect. But how does the technology that makes all this possible work?
Six years – an AI eternity
ChatGPT is a language model, i.e. a machine learning algorithm that specializes in processing texts. ChatGPT is the latest generation in a series of language models based on the so-called Transformer model introduced in 2017. The Transformer architecture caused a stir when it was released in professional circles because it enabled specialized language models for text translation and other tasks with unprecedented power.
Instead, the GPT model was pre-trained on very large data sets of generic texts in order to learn statistical properties of language as such independently of the specific task. The model prepared in this way could then be effectively adapted with smaller sets of sample data for specific tasks.
The next version GPT-2 appeared in 2019 (PDF) . It was essentially a scaled-up version of the previous model with a significantly higher number of parameters and with training on correspondingly larger data sets. In contrast to the original version, GPT-2 was no longer adapted for special problems, but was able to solve many different tasks such as translating texts or answering knowledge questions simply by training with generic texts from the Internet.
With 175 billion parameters, the third generation GPT-3 (PDF) was even more extensive than GPT-2 and correspondingly more powerful. It also attracted attention beyond the AI research community, particularly with its ability to write longer texts that were almost indistinguishable from those of human authors.
However, limitations of the model also became apparent, including ethical issues with objectionable or biased texts, and the habit of making grossly false statements of fact in persuasive-sounding language.
In order to remedy these shortcomings, OpenAI added a fundamentally new dimension to the training concept for its next language models InstructGPT and ChatGPT : Instead of leaving a model alone with huge amounts of text from the Internet, it was subsequently taught by human “teachers”, concrete ones To follow the instructions of the users and to make statements that are ethically justifiable and correct in terms of content. In order to ensure the effectiveness of this training, the algorithmic approach of the pure transformer model had to be expanded by a further step – the so-called reinforcement learning.
The impressive achievements of ChatGPT are the result of a whole range of different algorithms and methods as well as many tricks, some of which are very small. In this article, the focus is on providing an intuitive basic understanding of the technology without getting bogged down in too many mathematical or technical details. The links in the text refer to sources that fill in the gaps in this presentation.
