Day: January 10, 2024

2023 has been one of the most exciting years to witness the breakthrough of AI technology and Generative AI in particular, with the increasing popularity of ChatGPT (Generative Pretrained Transformer) and LLM (Large Language Models). This is thanks to its impressive ability to comprehend human languages and make decisions that remarkably mimic human intelligence.

ChatGPT reached an unprecedented milestone of 1 million users within five days. Since then, big tech giants have been quickly entering the race, releasing dozens of LLMs both open source and proprietary, such as LaMDA (Google AI), Megatron-Turing NLG (NVIDIA), PaLM (Google AI), Llama-2 (Meta AI), Bloom (Hugging Face), Wu Dao 2.0 (Beijing Academy of Artificial Intelligence), Jurassic-1 Jumbo (AI21 Labs) and Bard (Google AI), etc.

Alongside the race of big tech giants, the adoption of ChatGPT and LLMs in business is growing rapidly. According to the Master of Code Global report “Statistics of ChatGPT & Generative AI in business: 2023 Report”, 49 per cent of companies presently use ChatGPT, while 30 per cent intend to use it in the future.

Another report by Forbes suggests that 70 per cent of organisations are currently exploring generative AI, which includes LLMs. This suggests that LLMs are gaining traction in the enterprise world and that more and more companies are seeing the potential of this technology to revolutionise their businesses.

Multimodal Generative AI

Although ChatGPT and most other LLMs have been demonstrating superior performance in human language understanding (in text form), text is just one kind of data modal human beings perceive every day. However, multimodal data is ubiquitous in the real world, as humans often communicate and interact with all types of information, including images, audio, and video.

Multimodal data also poses significant challenges for artificial intelligence systems, such as data heterogeneity, data alignment, data fusion, data representation, model complexity, computational cost, and evaluation metrics. The AI community, therefore, often opts to successfully address the unimodal data before dealing with more challenging ones.

Inspired by the tremendous success of LLMs, the AI community has been creating Large Multimodal Models (LMMs) that can achieve similar levels of generality and expressiveness in the multimodal domain. LMMs can leverage massive amounts of multimodal data and perform diverse tasks with minimal supervision.

Also Read: Navigate in a cookie-less world, leverage AI and think community-first

Incorporating the other modalities into LLMs creates LMMs, which solve many challenging tasks involving text, images, audio, videos, etc., such as captioning images, visual question answering, and editing images by natural language commands etc.

OpenAI has been pioneering the development of GPT-4V, the upgraded multimodal version of the GPT-4 model that can understand and generate information from both text and image inputs. GPT-4V can perform various tasks, such as generating images from textual descriptions, answering questions about images, and editing images with natural language commands.

• LLaVA-1.5: This is a model that can understand and generate information from both text and images. It can perform tasks such as answering questions about images, generating captions for images, and editing images with natural language commands.
• Alpaca-LoRA: This is a model that can perform various natural language tasks by providing natural language instructions or prompts.

Adept, on the other hand, has been aiming at a bigger ambition: building an AI model that can interact with everything on your computer. “Adept is building an entirely new way to get things done. It takes your goals in plain language and turns them into actions on the software you use every day.” They believe that AI models reading and writing text are still valuable, but ones using computers like human beings are even more valuable to enterprise businesses.

This is driving the race among big tech companies to deliver Large Multimodal Models. It will take a few years for LMMs to reach the same levels as LLMs today.

Generating vs leveraging Large Foundation Models

Producing AI applications for many diverse tasks has never been easier and more efficient than before. Recalling several years ago, if we would like to make a sentiment analysis application, for example, it may take a few months to implement POC with both in-house and public datasets.

It also takes a few months to deploy the sentiment analysis models into the production system. Now, LLMs facilitate the development of such applications in a few days, simply formulating a prompt for LLMs to evaluate a text as positive, neutral, or negative.

In the field of computer vision, visual prompting techniques, introduced by Landing AI, also leverage the power of Large Vision Models (LVMs) to solve a variety of vision tasks, such as object detection, object recognition, semantic segmentation, etc.

Also Read: How Asia Pacific startups propel the evolution of Generative AI

Visual Prompting uses visual cues, such as images, icons, or patterns, to reprogram a pre-trained Large Vision Model for a new downstream task. Visual prompting can reduce the need for extensive data labelling and model training and enable faster and easier deployment of computer vision applications.

Generating pre-trained Large Foundation Models (LFMs), including LLMs and LVMs, requires not only AI expertise but also a huge investment in infrastructure, i.e., data lake and computing servers. Hence, the race to create pre-trained LFMs among big tech companies this year will continue in 2024 and in the years to come.

Some are proprietary, but many others are open source, leading to diverse alternatives for enterprises. Meanwhile, small and medium enterprises (SMEs) and AI startups will be the main forces in realising the commercials of LFMs. Thus, they will primarily focus on the creation of LFMs applications.

Agent concept in Generative AI

The agent concept is a new trend in Generative AI that has the potential to revolutionise the way we interact with computers. Agents are software modules that can autonomously or semi-autonomously spin up sessions (in this case, language models and other workflow-related sessions) as needed to pursue a goal.

One of the key benefits of using agents is that they can automate many of the tasks that are currently performed by humans. This can free up humans to focus on more strategic and creative tasks. Agents can be designed to be more user-friendly and easier to use than traditional Generative AI tools, making Generative AI more accessible to a wider range of users.

Here are some of the trends of agent concept in Generative AI:

• Increased use of agents to automate tasks: As Generative AI becomes more powerful and sophisticated, we can expect to see a greater use of agents to automate tasks that are currently performed by humans. For example, agents can be used to automate the process of creating and deploying AI models.
• Increased use of agents to make Generative AI more accessible: As agents become more user-friendly and easier to use, we can expect to see greater use of agents to make Generative AI more accessible to a wider range of users. This could lead to a new wave of innovation as more and more people are able to use Generative AI to create new products and services.
• Development of new agent-based Generative AI tools and platforms: As the agent concept becomes more popular, we can expect to see the development of new agent-based Generative AI tools and platforms. These tools and platforms will make it easier for developers to create and deploy agent-based Generative AI applications.

Also Read: Gen AI in banking: How to ensure a successful transformation for an age-old industry

Here are some specific examples of how the agent concept is being used in Generative AI today:

• Agent-based Generative AI tools: There are a number of agent-based Generative AI tools that are currently available. For example, Auto-GPT and BabyAGI are two tools that allow users to create and deploy agent-based Generative AI applications.
• Agent-based Generative AI platforms: There are also a number of agent-based Generative AI platforms that are currently available. For example, Google’s AI Platform and Amazon Web Services’ SageMaker platform both allow users to deploy and manage agent-based Generative AI applications.
• Agent-based Generative AI applications: There are a number of agent-based Generative AI applications that are currently in use. For example, agent-based Generative AI applications are being used to create new products and services, automate tasks, and make Generative AI more accessible to a wider range of users.

Overall, the agent concept is a new and promising trend in Generative AI. It is being used to develop new tools, platforms, and applications that are having a significant impact on a variety of industries.

—

Editor’s note: e27 aims to foster thought leadership by publishing views from the community. Share your opinion by submitting an article, video, podcast, or infographic

Join our e27 Telegram group, FB community, or like the e27 Facebook page

Image courtesy of the author

The post How AI will shape the future: A look ahead to 2024 and beyond appeared first on e27.

Silicon Box, an advanced semiconductor packaging company specialising in cutting-edge chiplet integration services, has secured US$200 million in a Series B fundraising round, bringing its valuation to over US$1 billion.

The capital came from the company founders, as well as other strategic investors, including BRV Capital, Event Horizon Capital, Maverick Capital, Prasedium Capital, Tata Electronics, TDK Ventures, and UMC Capital.

The money will be used to expand production in its advanced US$2-billion packaging factory in Singapore, which has been in mass production for early customers since October 2023, shortly after its grand opening on July 20, 2023.

Also Read: Semiconductor manufacturing nations set for growth as AI takes center stage: Alpha Intelligence Capital CEO

Current semiconductor chips are hitting a wall in scalability, limited by conventional packaging approaches. Meanwhile, chip designers’ development and manufacturing costs have become cost-prohibitive except for the most well-funded players, leaving the industry bottlenecked and consumers paying high prices.

Founded in 2021 by semiconductor design and packaging industry titans Dr Sehat Sutardja and Weili Dai and CEO Dr Byung Joon Han, Silicon Box aims to bring affordable, high-performance, power-optimised, scalable solutions that enable next-gen large language models (LLM), generative AI, automotive, data centres and mobile computing.

The startup enables chiplet architecture, allowing chip designers freedom from the constraints of a single, monolithic chip for processing. By leveraging multiple smaller chips interconnected in a single package, chip designers can create the equivalent of a “system-on-a-chip” (SoC) in a package.

Chiplets enable dramatically better performance, smaller device sizes, and better device reliability. Most importantly, they make it easier for foundries and chip designers to collaborate to build chips for the most cutting-edge applications.

The company claims its solutions are more reliable and cost-effective due to the standardised packaging process for the shortest chiplet-to-chiplet interconnection, reducing the manufacturing costs for high-performance devices by up to 90 per cent, with better thermal and electrical performance. This is especially crucial for the high-growth AI accelerator market.

Also Read: How Infineon aims to build better semiconductors with the help of Singapore startups

“We are leading the pack to bring high performance, power-optimised, affordable, and scalable solutions that enable next-gen large language models (LLM), generative AI, automotive, data centres, and mobile computing globally,” said CEO Joon Han. “Our state-of-the-art factory and advanced panel-level packaging are delivering a solution to scale high-growth markets, such as AI accelerators, to the masses.”

The post Silicon Box bags US$200M to expand production in its US$2B semiconductor factory in Singapore appeared first on e27.

Indonesia-based HR tech company VENTENY today announced that it has acquired 30 per cent of shares in PT Digitalisasi Perangkat Indonesia (DPI), a fintech holding company providing integrated digital banking services which includes rural digitalisation, financing, e-money, and e-remittance.

In a press statement, the company said that the acquisition is a way for VENTENY to acquire DPI’s e-money license, which it has secured for its Dipay platform.

The integration of Dipay into the VENTENY ecosystem will allow users to pay bills, top-up credits, and send remittances. It is also expected to increase the level of security in the ecosystem.

In the future, VENTENY also wants to introduce new features, such as enabling users to pay employee salaries using the e-money feature.

Also Read: How AnyMind Group achieved profitability through its approach to human resource and leadership

“Through this corporate action, VENTENY aims to optimise the integration of digital payment technology to help users perform transactions. In addition to giving added value for users, the e-money integration will enable a more practical and efficient payment experience through the VENTENY Employee Super App,” said VENTENY Founder and Group CEO Jun Waide.

IDX-listed VENTENY provides an “employee super app” ecosystem with features including insurance, employee benefits, and corporate training for small- and medium-sized enterprises (SMEs).

The company operates in the Philippines, Singapore, Indonesia, and Japan. It claimed 9,600 SMEs onboard its platform, with over 250,000 users of its app in Indonesia.

VENTENY said it recorded a 125 per cent increase in revenue in Q3 2023 compared to Q3 2022. It has also experienced a 57 per cent increase in profits from Q2 2022 to Q3 2023.

—

Image Credit: RunwayML

The post VENTENY acquires 30% of Dipay’s parent company to integrate e-money offering into its app appeared first on e27.

Dear reader,

VinFast’s recent agreement to build its first electric vehicle (EV) facilities in India marks a significant stride for the Vietnamese automaker. The US$2 billion investment in the South Indian state of Tamil Nadu signals the car maker’s ambitious foray into the world’s third-largest vehicle market.

With an initial commitment of US$500 million over five years, the company aims to commence construction later this year, creating 3,000 to 3,500 local job opportunities. In Thoothukudi, the facility plans to manufacture EV batteries, anticipating an annual capacity of up to 150,000 vehicles.

While this falls short of its Vietnam plant’s capacity, VinFast envisions Tamil Nadu evolving into a pivotal EV production hub. As India targets 30% electric vehicle sales by 2030, VinFast’s strategic move aligns with the government’s initiatives.

The success of this venture could not only propel VinFast’s global presence but also contribute significantly to India’s burgeoning electric vehicle landscape.

Sainul,
Editor.

====

Asia Partners hits final close of Fund II at US$474M
The LPS include IDFC, FIC, DEG, and Generation Capital; Asia Partners, which typically invests between US$20M and US$100M, has already backed companies such as ShopBack and Doctor Anywhere.

Vietnam’s VinFast to set up US$500M EV facilities in India
VinFast and the southern state of Tamil Nadu agreed to work toward an investment of up to US$2B; Construction on the project’s EV and battery manufacturing plants is expected to start this year and generate 3,000-3,500 jobs locally.

Silicon Box bags US$200M for its US$2B semiconductor factory in Singapore
Investors include BRV Capital, Maverick Capital, Prasedium Capital, and Tata Electronics; Silicon Box is an advanced semiconductor packaging company specialising in cutting-edge chiplet integration services.

Atome raises US$31M from parent firm Advance Intelligence Group
The latest capital infusion comes just months after the buy now, pay later platform renewed its U$100M debt facility from HSBC Singapore to help Atome expand in the Philippines.

Komunal lands US$5.5M in Series A+ round to digitalise rural banks in Indonesia
Investors include Sumitomo Corporation, Jafco Asia, Skystar Capital, and Gobi Partners; Komunal offers neo-rural bank services and has so far partnered with 376 rural banks and channels productive loans to MSMEs based in tier 2 and 3 cities.

Ant Group close to buying Dutch payments firm MultiSafepay for US$200M
The acquisition follows Ant Group’s purchase of Singapore-based payments firm 2C2P in 2022; Ant, which operates cross-border payments platform Alipay+, also bought British payments group WorldFirst in 2019 for US$700M.

Salarium halts operations amid salary disbursement trouble
In an email sent to clients, Salarium mentioned “issues affecting its ability to continue operating effectively,” including an incident happening on December 18; It had “until now been unable to be resolved,” leading to the shutdown.

Semaai nets US$4.7M to expand its agritech, fintech solutions to Central Java
Investors are CyberAgent Capital, Sumitomo Corporation, Ruvento, MyAsiaVC, Heracles Ventures, and Beenext; Semaai helps farmers and rural MSMEs in Indonesia maximise their earning potential and access better financing, services and new markets.

TikTok steps up US e-commerce push with ambitious sales goal
TikTok’s ambitious goal will see it push harder to redirect users’ attention from short videos to in-app shopping, a potential threat to established US e-commerce giant Amazon.

Volkswagen is bringing ChatGPT into its cars and SUVs
Volkswagen has integrated ChatGPT into the backend of its IDA voice assistant, which drivers can use to control the infotainment, navigation and air conditioning, or to answer general knowledge.

North Korean hackers stole US$600M in crypto in 2023: Report
With nearly US$1.5B stolen in the last two years alone, North Korea’s hacking prowess necessitates constant vigilance and innovation on the part of businesses and governments, the report suggested.

Climate tech funding drops 30% in 2023
Despite a notable peak in financing in the third quarter of the year, which reeled in US$12.5B, funding in 2023 was in a notable decline compared to the previous years since 2020; Deal counts also decreased for the first time since 2020, with figures down 3% y-o-y.

US court orders Nasdaq-listed Society Pass to pay US$1.1M to ex-CTO
Rahul Narain accused Society Pass of failing to honour the contract terms and pay him his earned salary compensation, bonus payments, healthcare reimbursements, equity awards, and severance pay.

What founders need to know about creating a cap table
When preparing for a cap table, pre- and post-money valuations are some of the key elements that founders must consider and include; One of those mistakes is forgetting to add ESOPs, which will lead to a piece of missing info that can affect the accuracy of the calculations.

What 3 things should Web3 founders think about in 2024?
Experts from Cointelegraph’s Accelerator agree: Blockchain developers should prioritise interoperability if they want to achieve greater adoption for Web3.

From revolution to real-world value: How companies can benefit from Web3 in 2024
Web3 is valuable when focused on specific scenarios where it adds to what Web2 alone can already offer; The best examples of Web3 apps moving from general-purpose to specific technologies are in the realm of tokenisation.

GenAI could make KYC effectively useless
Viral posts on X (formerly Twitter) and Reddit show how, leveraging open source and off-the-shelf software, an attacker could download a selfie of a person, edit it with generative AI tools and use the manipulated ID image to pass a KYC test.

How young D2C brands are using AI to transform customer growth and retention
Velocity matters when scaling a D2C e-commerce brand, and success is about the number of smart bets you can take quickly.

How AI will shape the future: A look ahead to 2024 and beyond
2023 has been one of the most exciting years to witness the breakthrough of AI technology and Generative AI in particular.

The canary in Singapore’s retail coal mine is ‘kiasu’
While the average Singaporean is even more picky than usual this holiday season, adjusting their spending to combat the rising cost of living, high inflation, and a looming recession, under these economic challenges, the kiasu shopper is more deal-hungry than ever.

Navigating the regulatory landscape: Malaysia’s startup outlook in 2024
The Securities Commission of Malaysia in December 2023 said that it is seeking to introduce a “small offering exemption” in the current securities law in 2024; The exemption may permit safe harbour for offerings of a “certain size” to sophisticated investors.

==

The post Ecosystem Roundup: VinFast to invest US$500M in India; Asia Partners closes US$474M Fund II; Silicon Box bags US$200M appeared first on e27.

South Korea’s early-stage startup investor, Infobank Corporation, has partnered with Singapore’s Farquhar VC (FVC) to help Korean startups succeed globally.

Under this partnership, iAccel, the investment department of Infobank, and Farquhar will collaborate comprehensively on exchanging innovation and investment ecosystem insights, mutually supporting each other’s portfolio companies and jointly participating in bids for global acceleration initiatives.

Also Read: Farquhar VC to help Korean university-affiliated startups to go global

Both parties will also explore potential opportunities for joint funding that could support Korean startups seeking global expansion and global startups seeking to establish a foothold in South Korea.

Infobank iAccel Deputy CEO Gil Chang-Gun said that both parties would commence their cooperation via potential co-investment into data-centric Infobank portfolio companies that are advancing into Southeast Asia. Also, the Infobank-FVC partnership shall strengthen the ecosystem between North Asia and Southeast Asia.

As of November 2023, iAccel has invested more than KRW 20 billion (US$150 million) into more than 80 companies. Two of its portfolio companies (Crowdworks and Qualitas Semiconductor) have made IPO debuts on KOSDAQ.

Established in 2020, Farquhar VC has invested in nearly 40 startups to date, with Lomotif as one of its early exits. It also works closely with early-stage startups to support their growth through targeted market access with its mid-sized and large enterprises network.

Also Read: Singapore’s Farquhar VC joins StockViva’s US$5M Series A investment round

In particular, FVC’s innovation advisory arm recently undertook an accelerator programme on behalf of the Seoul Business Agency, which garnered three cross-border strategic partnerships and one investment commitment.

—

The post Infobank, Farquhar to help Korean startups seeking global expansion appeared first on e27.