• Posts
  • SoftBank set to back AI start-up Perplexity

SoftBank set to back AI start-up Perplexity

Also: Meta starts testing user-created AI chatbots on Instagram


Today’s newsletter highlights key developments in the AI sector. SoftBank plans a substantial investment in AI start-up Perplexity, potentially boosting its valuation to $3 billion, despite ongoing plagiarism accusations. Meta has begun testing user-created AI chatbots on Instagram to enhance user engagement. Goldman Sachs has deployed its first generative AI tool across the firm, significantly increasing developer productivity. A critical look at the impact of AI on the freelance job market reveals growing challenges for gig workers. Additionally, efforts to ground enterprise AI in factual data are intensifying, with companies like KPMG and OpenAI leading the charge. In the venture capital space, Europe has set a new record for generative AI investments, reflecting the region’s growing commitment to AI innovation.


  • 💵 SoftBank set to back AI start-up Perplexity

  • 👩‍🦳 Meta starts testing user-created AI chatbots on Instagram

  • 🏦 Goldman Sachs Deploys Its First Generative AI Tool Across the Firm

  • ⚠️ Is AI killing freelance jobs?

  • 👩🏻‍💻 The AI industry is working hard to ‘ground’ enterprise AI in fact

SoftBank plans to invest $10 million to $20 million in the AI start-up Perplexity, which could elevate the company’s valuation to $3 billion. Founded in August 2022, Perplexity is known for its generative AI platform with a chatbot interface similar to ChatGPT and Microsoft’s Bing. The start-up has previously raised significant funds, including $25.6 million in Series A and $73.6 million in Series B funding from prominent investors like Nvidia and Jeff Bezos. Despite its success, Perplexity faces plagiarism accusations, notably from Forbes, and criticism from Wired for allegedly bypassing web standards to scrape content. Perplexity co-founder and CEO Aravind Srinivas contends that their platform aggregates information and denies intentional content theft.

Meta is testing user-created AI chatbots on Instagram, enabling creators to design AI characters via Meta AI studio. This initiative, starting in the U.S., aims to enhance user engagement by introducing AI avatars in messaging, labeled clearly as AI. Early versions involve collaborations with creators like Wasted and Don Allen Stevenson III. Mark Zuckerberg highlighted the potential for these AI chatbots to aid creators and businesses in interacting with their communities and customers, addressing the challenge of managing high volumes of messages. The project is in its early stages, with Meta focusing on improving the AI’s effectiveness and expanding access over time.

Goldman Sachs has launched its first generative AI tool, aimed at enhancing code generation, for thousands of developers within the firm. This initiative is part of a broader strategy to centralize all AI technologies on an internal platform, known as the GS AI Platform, which incorporates models like GPT-3.5, GPT-4, and Meta’s Llama, among others. This centralized approach, while initially slowing progress, has resulted in significant efficiency gains, including a 20% increase in developer productivity. The platform supports various applications, such as a copilot tool for investment bankers, and is designed to ensure data security and regulatory compliance. This deployment reflects a cautious but strategic integration of AI to boost operational efficiency without compromising safety and compliance.

Advancements in AI, particularly generative AI tools like ChatGPT and Copilot, are significantly impacting the freelance job market by reducing the availability of roles traditionally filled by freelancers. Jobs such as copywriting and coding, which have been mainstays of freelance work, are now increasingly automated. This shift is making freelancers, who already face limited job security, particularly vulnerable. Data indicates that many roles historically suitable for freelancers are among the most susceptible to AI replacement, pushing gig workers to consider pivoting to new types of work. The full extent of this transition and the future opportunities for these workers remain uncertain.

The AI industry is intensifying efforts to ground enterprise AI models in factual data to address the issue of generative AI tools producing false information, or “hallucinations.” Companies are increasingly focusing on integrating AI models with their proprietary data to ensure accuracy and reliability. For instance, KPMG and Microsoft Azure have developed an AI search tool that accesses corporate data without replicating it, while OpenAI’s acquisition of Rockset aims to enhance real-time data processing capabilities. Additionally, Google’s Vertex AI now incorporates data from reliable third-party sources such as Moody’s and Thomson Reuters. These measures are part of a broader strategy to make AI tools more effective and trustworthy for enterprise applications, ensuring that businesses can leverage AI without compromising on data integrity.

🛠️ AI tools updates

Synthesia has launched Synthesia 2.0, a significant upgrade to its AI video creation platform, aimed at enhancing business video production. This new version introduces advanced features such as improved video quality, customizable avatars, and more streamlined editing tools. Businesses can now create professional-grade videos more efficiently, leveraging AI to handle various aspects of production from script to final edit. The platform’s enhancements are designed to make video creation more accessible and less time-consuming, enabling companies to produce high-quality video content for training, marketing, and internal communications without needing extensive video production expertise.

💵 Venture Capital updates

Europe has achieved a new record in venture capital investment for generative AI startups, with €1.9 billion ($2 billion) invested in the first half of 2024 across 61 funding rounds, surpassing last year’s total of €1.6 billion. Significant deals include €600 million for Paris-based Mistral AI and $300 million for German AI translation startup DeepL. France leads in deal value, driven by strong university programs and government support, while the UK boasts the highest deal count. Despite these achievements, Europe’s generative AI funding remains behind the US, which has invested €15.9 billion in the sector. The European Commission’s ongoing investment of €1 billion annually in AI and plans to increase this to €20 billion per year by the end of the decade underscore the region’s commitment to maintaining its growth trajectory.

🫡 Meme of the day

⭐️ Generative AI image of the day