TLDR

Alibaba: Awaiting the Rebound

In recent years, Alibaba has faced a series of challenges that have left the company on the back foot. The Chinese government's push for shared prosperity, combined with the broader economic slowdown and rising competition from platforms like Temu and Shein, has raised questions about the future of what was once a dominant tech giant in the East.

However, there are signs that Alibaba is finding new ways to pivot. In November, alongside its famed Single’s Day shopping event, the company introduced Qwen 2.5-Coder, an open-source LLM (large language model) that rivals the performance of GPT-4o and Claude 3.5 Sonnet. This move marks a significant shift away from Alibaba’s core e-commerce focus, signaling that the company is looking to carve out a new battleground in the rapidly expanding AI space.

This edition aims to take a deeper look at Alibaba’s AI strategy in 2024 and onward.

Qwen 2.5-Coder: A Game Changer?

For those not yet familiar with Qwen or its evolving role in AI, let’s start with a brief overview of what it brings to the table:

• Performance: Qwen 2.5-Coder delivers solid performance on key benchmarks, rivaling other major coding models. While it excels at algorithmic tasks and problem-solving, it still lags behind in advanced error handling and full-stack development. Simply put, it’s an affordable, high-performing option for developers.

• Open source: The real appeal of Qwen lies in its open-source nature. According to a well-regarded AI tech article, the Qwen series has spawned over 50,000 derivative models, second only to Meta’s Llama series. This widespread adoption has turned Qwen into one of the most influential open-source models globally, making AI development more accessible to smaller players who can't compete with the vast resources of giants like Google or Microsoft.

• Local execution: Qwen 2.5-Coder 32B can be run entirely locally, allowing users to avoid hefty API fees. Alibaba has also released optimized versions of the model, reducing file sizes to make them more accessible for users with less powerful hardware. For instance, the Q2 quantized version is only 12 GB, down from the original 20 GB.

Alibaba’s Strategic Shift: AI Over E-commerce?

Unlike competitors like Baidu and Huawei, which have focused their LLMs on specialized sectors like autonomous driving and hardware, Alibaba is making a bold move to make AI more accessible to everyone. The company’s motto, “AI for all,” reflects its commitment to the open-source ecosystem and its broad vision of driving technological innovation.

Cloud and AI Synergy

Alibaba’s shift from e-commerce to cloud services plays a crucial role in its AI strategy. Tongyi Qianwen, Alibaba's flagship LLM, powers everything from content creation to customer support. Its deep integration across Alibaba's cloud infrastructure offers unparalleled access to data, which is key to continuously improving AI models.

In fact, 80% of Chinese tech companies and half of the leading LLM companies rely on Alibaba Cloud, a scale unmatched by any competitor. The company has also created ModelScope, the largest open-source AI community, offering a plethora of AI models to developers around the world.

Making Qwen Front and Center

Alibaba is rapidly deploying AI across its products, a strategy that mirrors the moves of major Western tech firms like Google and Microsoft. Some examples include:

• Tongyi - AI reading assistant

• Renxin - AI medical assistant

• Farui - AI legal advisor

• Tingwu - AI audio transcriber

• Wanxiang - AI image generator

• Lingma - Al coder

• Nichang - AI virtual dresser based on Chinese history and culture

• Chartcube - AI chart generator

• Imgcook - AI design-to-code generator

• Ihuiwa - AI E-commerce assistant for content generation

• The list continues…

Today, Alibaba uses AI chatbots to handle customer engagements for more than two million daily sessions and over 10 million lines of daily conversations on Taobao’s two-sided platform, representing about 75% of Alibaba’s online and 40% of phone hotline consultations. Driving cost reduction by over $150M while improving satisfaction by 25%.

The detail integration as below:

• Wanxiang-bot: Designed to support merchants on the Taobao platform, the Wanxiang-bot helps resolve issues related to platform rules, promotional activities, and general service inquiries.

  Alibee Shop Bot: This bot focuses on improving merchant-to-consumer interactions. It assists merchants by handling customer service issues and facilitates seamless communication between merchants and end users, ensuring smooth transactions on Taobao.

  Alime Bot: Aimed at the end user, Alime helps consumers by providing support through both online and phone hotline channels. It offers a rich array of interactive features—text dialogues, visual cards, images, videos, and other dynamic elements—that enhance the user experience during conversations.

  AI-bot: Working alongside Alime, the AI-bot steps in to handle service disputes between consumers and merchants. It uses algorithms to analyze transaction data and previous conversations to assess the situation and make automated judgments. The AI-bot then reaches out to consumers to communicate the dispute resolution. If the consumer is satisfied with the outcome, the case is closed. If not, the AI-bot assists with filing an appeal, which will be reviewed by a human evaluator.

  Dahuang-bot: This bot is designed to enhance customer service training by simulating a wider variety of consumer and merchant interactions than traditional, non-AI-based training systems. By creating a more diverse and extensive training environment, Dahuang-bot helps improve the effectiveness of customer service teams.

Conclusion: A Bold Transformation

Alibaba’s AI investments go beyond software; the company is also actively developing its own AI hardware. Through T-Head, its chip business, Alibaba is pursuing semiconductor self-sufficiency, which is especially important given the current geopolitical climate.

By investing in AI companies like Moonshot AI, Baichuan, and MiniMax (the “Big 4 AI Tigers”), Alibaba is positioning itself at the forefront of AI development, both in terms of software and hardware.

This is a bold, even desperate, move towards the AI phenomenon, but it’s also one that is cohesive, directional and also driving in face of an unprecedented challenge. Not all face this the same way. Though audacious, but yet prudent. Isn’t this one of the qualities we are looking for from founder mode?

What do you think? Will Alibaba’s bold AI strategy help it reclaim its former glory, or is it too late for the company to turn things around?

Let me handle it

What is this?

Windsurf is an AI-powered in-browser code editor by Codeium that offers real-time autocompletion, debugging, and collaboration features for seamless programming workflows.

What can you use it for?

Coding. While Cursor has a higher reputation in the domain because of its earlier entrance, and a little girl helping it to promote, Windsurf has a lower entry barrier in comparison and can make things just a bit more comfortable if you don’t have development background.

You can learn more from the video below:

What’s the price?

It’s free, plus a monthly subscription if you wish to access more powerful coding models and advanced features.

Pro tips

• Since the edition is all about a coding model, what would you guess? Remember to link it with Qwen and give it a try!

What’s new in town

Food of Thought

Why so serious?