Author: yasmeeta
Secoda secures $14 million in funding to introduce AI-powered, Google-inspired search capabilities for corporate data.
Secoda, a Toronto-based company specializing in AI-powered data search, cataloging, lineage, and documentation, has recently announced a successful Series A funding round, securing $14 million in investment. This capital infusion is earmarked for the advancement of its AI solutions, with the ultimate goal of enabling any enterprise user, regardless of their technical expertise, to seamlessly access, comprehend, and utilize company data. The user experience is designed to be as effortless as conducting a Google search.
This funding round brings Secoda’s total funding to $16 million and is spearheaded by its existing investor, Craft Ventures, with notable participation from Abstract Ventures, YCombinator, and Garage Capital. Distinguished figures from the data ecosystem, including Jordan Tigani (CEO of MotherDuck), Scott Breitenother (CEO of Brooklyn Data), and Tristan Handy (CEO of dbt), have also joined in this round.
Jeff Fluhr, co-founder and partner at Craft Ventures, emphasized the growing significance of data lineage understanding and data utilization for companies. He stated, “It has become increasingly important that companies not only have a full understanding of the lineage of their data from disparate sources but also harness their data to make more efficient and informed decisions. Secoda has built a powerful AI-powered data copilot for companies to do just that.”
The Current Data Challenge:
In today’s enterprise IT landscape, numerous systems are designed for diverse functions, creating a complex web of technologies that are essential for organizational efficiency. However, this complexity often results in disconnected data sources, with applications failing to communicate effectively and data remaining isolated.
Consequently, employees face challenges when seeking answers related to data. They must navigate through convoluted applications or seek assistance from the data team, diverting attention from other tasks. Etai Mizrahi, who encountered similar issues while working at Acadium, described the frustration: “Questions that seem simple enough to answer end up feeling like a huge, frustrating game of broken telephone.”
Addressing the Knowledge Gap: To bridge this knowledge gap, Mizrahi collaborated with colleague Andrew McEwen to launch Secoda in 2021, offering an all-in-one platform for data management and search. Secoda seamlessly integrates with business intelligence and transformation tools, as well as data warehouses, effectively connecting all components of a team’s fragmented tech stack to establish a single source of truth for company data.
To simplify the process further, Secoda employs a ChatGPT-powered assistant, enabling users to create documentation that adds contextual information to metadata. Users can also search their company’s consolidated data catalog using natural language queries. Mizrahi highlighted the platform’s capabilities, stating, “Secoda does not simply give you information but gives you answers, much like Google… Our customers have been able to leverage the platform to reduce the volume of inbound data requests by over 40%, reduce onboarding times by 50%, and reduce time teams spend on documentation by 90% — huge time savings for data teams.”
Future Plans:
With the recent funding injection, Secoda plans to bolster its engineering team and intensify research and development efforts, particularly in the area of AI enhancements. Additionally, the company intends to introduce “Secoda Monitoring” to assist data teams in ensuring the quality and accuracy of the data they use. This feature will provide insights into how changes affect assets and reduce data quality errors. Monitoring will also enable companies to track the operational efficiency of their data teams and identify potential cost savings.
Over the past year, Secoda has experienced substantial growth, expanding its customer base by a factor of five and managing over 100 million metadata resources, including tables, dashboards, columns, and queries. On the integration front, the data search tool currently supports 36 popular data warehouses, business intelligence tools, and productivity platforms, including Snowflake, dbt, and Looker. The company is committed to continually adding more connectors based on popularity and user demand.
Voyager Capital aims to raise $100M for its sixth fund to support Pacific Northwest B2B startups.
Voyager Capital is in the process of securing funds for yet another investment fund aimed at supporting early-stage business-to-business startups situated in the Pacific Northwest.
This latest development comes to light through a recently filed document with the SEC, marking this as Voyager’s sixth fund since its inception 25 years ago.
Diane Fraiman, the Managing Director at Voyager, confirmed this news following an inquiry from GeekWire. The goal for this sixth fund is to amass $100 million, with a “hard cap” set at $125 million, according to Fraiman.
Notably, Voyager had previously raised $100 million for its fifth fund back in 2019, following an earlier $50 million fund in 2013.
Established in 1997, Voyager has provided support to over 75 companies within the Pacific Northwest, accumulating assets under management exceeding $520 million.
With offices in Seattle and Portland, Voyager concentrates its investments in startups hailing from Washington, Oregon, British Columbia, and Alberta. These startups primarily focus on developing software-as-a-service, cloud infrastructure, artificial intelligence, and machine learning-related products and services.
Some of the companies currently within Voyager’s portfolio include Carbon Robotics, Treasury4, WellSaid Labs, Hiya, Syndio, among others. Notable exits include Zipwhip, acquired by Twilio in 2021, and Yapta, acquired by Coupa Software in 2020.
The venture capital landscape has faced challenges recently due to higher interest rates and a tech market slowdown. PitchBook reported a total of $33.3 billion raised across 233 venture funds in the first half of the year, compared to over $167 billion raised in 2022.
During a panel discussion in downtown Seattle earlier this summer, Bill McAleer, Managing Director at Voyager Capital, offered insights into factors he believes will rejuvenate the venture capital market. These factors include tech professionals laid off from larger companies seeking opportunities in startups, the revaluation of early-stage companies offering advantages to venture capitalists, and the growing adoption of generative AI, with McAleer predicting that the AI tools market will soon rival or even surpass the impact of cloud technology.
Several other Pacific Northwest firms have also announced new funds this year, including PSL Ventures, Ascend, Madrona Venture Labs, and AI2.
Google Bard now has the capability to seamlessly access Gmail, Docs, Maps, and other services
Today, Google unveiled a significant upgrade to its Bard conversational AI system, broadening its functionality to seamlessly interact with Google’s most popular productivity apps and services. These improvements are designed to enhance Bard’s utility in daily tasks while also addressing concerns regarding its accuracy.
Effective immediately, Bard gains the ability to directly tap into information from apps such as Gmail, Docs, Maps, Flights, and YouTube. This empowers it to deliver more comprehensive and personalized responses during conversations. For instance, when planning a trip, Bard can now autonomously retrieve pertinent dates, flight details, directions, and sightseeing recommendations, all within a single conversation.
This upgrade follows Bard’s somewhat lackluster public debut in March, which exposed factual inaccuracies in many of its responses. Google aims to bolster Bard’s accuracy by integrating it with its search engine. Users now have the option to fact-check Bard’s responses against web information indexed by clicking a “Google it” button within the chat.
Google emphasizes its unwavering commitment to safeguarding users’ personal information with this update. Those who choose to utilize the Workspace extensions can rest assured that their content from Gmail, Docs, and Drive remains confidential, unseen by human reviewers, not utilized for ad targeting by Bard, nor employed in training the Bard model. Users retain full control over their privacy settings.
Additionally, Google has simplified the process of building upon others’ interactions with Bard. From today onwards, if someone shares a Bard chat via a public link, the recipient can continue the conversation, pose further inquiries to Bard about the topic, or use it as a starting point for their own discussions.
Furthermore, Google extends access to existing English language features, including the ability to upload images using Lens, receive Search images in responses, and adapt Bard’s responses to over 40 languages.
Although Bard’s capabilities are still somewhat limited, these new features hint at a future where AI assistants seamlessly combine conversational skills with email and document services to enhance productivity. As Bard continues to advance, it may integrate further with other Google offerings, such as calendar, photos, and analytics.
The rollout of these Bard upgrades begins today, with Google planning to introduce additional languages and integrations in the months ahead, all while upholding responsible technology refinement.
HiddenLayer secures $50 million in funding to strengthen the security of corporate AI models.
HiddenLayer, a cybersecurity startup based in Austin, Texas, emerged in response to a cyberattack that exploited machine learning code at the founders’ previous company. Today, HiddenLayer has announced a successful $50 million Series A funding round aimed at bolstering the defenses of the ever-expanding array of AI models adopted by enterprises.
This funding round was spearheaded by M12, Microsoft’s Venture Fund, and Moore Strategic Ventures, with participation from Booz Allen Ventures, IBM Ventures, Capital One Ventures, and Ten Eleven Ventures.
HiddenLayer’s CEO and Co-Founder, Chris Sestito, expressed, “The rapid adoption of AI inspires us to accelerate our mission, ensuring that every security professional possesses the necessary tools and expertise to embrace AI securely.”
HiddenLayer already plays a crucial role in safeguarding AI/ML models for numerous Fortune 100 companies across various sectors, including finance, government, defense, and cybersecurity.
What HiddenLayer Offers:
Previously covered by VentureBeat, HiddenLayer has developed a suite of tools as part of its “MLSec” Platform, designed to safeguard enterprise machine learning (ML) and AI models. These tools do not directly access the models or compromise clients’ proprietary data and technology. Instead, they actively monitor the performance and operations of enterprise ML/AI models and associated applications in real-time. They scan for overarching vulnerabilities, provide recommendations for fortifying security, and detect the injection of malicious code/malware. Furthermore, they deploy defense mechanisms to thwart attackers and isolate intrusions.
HiddenLayer’s MLSec Platform includes an intuitive yet powerful dashboard, granting security managers immediate access to essential information about the security status of their enterprise ML/AI models. It automatically lists security issues and alerts in order of priority based on severity and stores data for compliance, auditing, and reporting purposes.
Additionally, HiddenLayer offers consulting services provided by their team of Adversarial Machine Learning (AML) experts, who remain up-to-date with the latest security trends and threats. These services encompass threat assessments, training for cybersecurity and DevOps personnel, and “red team” exercises to verify the effectiveness of clients’ defenses.
Key Partnership:
Earlier this year, HiddenLayer established a partnership with the prominent enterprise data lakehouse provider, Databricks. This collaboration enables Databricks enterprise customers to directly utilize HiddenLayer’s MLSec Platform for their models running on Databricks’ lakehouses. The integration is model-agnostic and covers model scanning, detection, and response. It empowers Data Scientists and ML Engineers to enhance their models’ security without altering their code or environment. As models are loaded, HiddenLayer’s model scanner ensures integrity and security. If an attack is detected, the integration responds automatically without human intervention.
Future Goals for Enterprise AI Security:
HiddenLayer’s inception traces back to an incident at its co-founders’ prior company, Cylance, where ML models fell victim to a cyberattack. Attackers exploited Cylance’s Windows executable ML model, revealing vulnerabilities and enabling the production of evasive binary files. While this event was concerning, it prompted the realization that attacks on ML/AI would escalate as more enterprises adopted generative AI to boost efficiency and performance.
Presently, HiddenLayer is experiencing rapid growth, having quadrupled its workforce in the past year. With the Series A funding secured, the company plans to hire an additional 40 personnel by year-end and continue expanding its client base.
Step aside, AI; quantum computing is poised to become the most formidable and unsettling technology.
In 2022, leaders within the U.S. military technology and cybersecurity community expressed their view that 2023 would serve as the pivotal “reset year” for quantum computing. They made an estimation that aligns the timeline for securing systems against quantum threats with the emergence of the first quantum computers capable of jeopardizing security, both anticipated within a span of approximately four to six years. It is of paramount importance that industry leaders swiftly grasp the security implications of quantum computing and take proactive measures to address the challenges poised by this formidable technology.
Quantum computing stands at the forefront of technological innovation, presenting an unparalleled array of challenges while holding the promise of unprecedented computational prowess. Unlike conventional computing, which relies on binary logic (comprising 0s and 1s) and sequential operations, quantum computing harnesses the power of quantum bits, or qubits, capable of representing an infinite spectrum of potential outcomes. This characteristic empowers quantum computers to execute an immensely large number of calculations concurrently, exploiting the probabilistic nature of quantum mechanics.
The Potential of Quantum Computing
The potential inherent in quantum computing lies in its capacity to process vast volumes of data in parallel, resulting in exponential leaps in computational capabilities when compared to classical computers. Whereas a classical computer can determine the outcome of a single-person race, a quantum computer can simultaneously analyze a race involving millions of participants with diverse routes and ascertain the most probable victor through probability-based algorithms. Quantum computers are exceptionally well-suited to tackle optimization problems and simulations featuring multiple probabilistic outcomes, thereby revolutionizing domains such as logistics, healthcare, finance, cybersecurity, weather prediction, agriculture, and more. Their influence could extend to the realm of geopolitics, fundamentally reshaping global power dynamics.
Quantum computing necessitates a fundamentally distinct approach to programming due to its novel logical framework. Embracing uncertainty and iterative heuristic methods are imperative for effectively harnessing the potential of this technology. Nevertheless, a substantial hurdle in the field of quantum computing is the challenge of connecting multiple qubits without elevating the likelihood of errors, which remains a critical impediment to the commercial growth of the technology.
A Practical Constraint and Ongoing Research
One practical limitation involves the imperative to shield qubits from the real-world environment to prevent decoherence, which can degrade the quantum state. Currently, achieving isolation involves cooling to extremely low temperatures. Ongoing research endeavors explore diverse methodologies, including photonics and various materials, with the goal of rendering quantum processors more scalable and commercially viable.
A Thousand-Qubit Milestone
Over the past decade, quantum computing has achieved remarkable progress. For instance, in 2017, IBM unveiled a 50-qubit chip, and in 2019, it claimed to have surpassed the fastest traditional supercomputer in specific computations. Further advancements are anticipated, with the race to develop 1,000-qubit quantum computers already in progress.
While short-term expectations regarding quantum computing may sometimes be overblown, the long-term implications are poised to be transformative. Increasing global interest across various sectors is accompanied by substantial financial commitments, laying the groundwork for extraordinary practical innovations in the years ahead.
Crucial Development of Error-Correcting Qubits
To fully unlock the potential of quantum computing, the development of error-correcting qubits assumes critical importance. Presently, quantum processors often necessitate a significant number of standard qubits to produce a single error-correcting qubit. Nevertheless, there is optimism within the community that this challenge will be effectively addressed in the coming years.
Quantum Computing’s Promise
Quantum computing holds the promise of reshaping our world by bestowing upon us unparalleled computational capabilities and revolutionizing diverse industries and fields. While challenges persist, the persistent progress in quantum technology suggests that breakthroughs could materialize at any moment. As we harness the potential of quantum computing, it is poised to emerge as the most influential of all frontier technologies, propelling significant advancements within our society.