Sign up for our newsletter.

times popup

Month: August 2024

The Changing Landscape of Data in the AI Era

Posted on by lucy lee

Data has always been the lifeblood of business. In the era of artificial intelligence, it is all the more important. The quality and quantity of data directly determines the performance of AI models and the insights they generate. As AI becomes more integrated across industries, data pipelines and architectures are rapidly evolving to meet these new challenges. In this article, we’ll delve into how data companies are adapting and the opportunities that lie ahead.

The Primacy of Data in AI

In the past, there’s been a great focus on the quality and complexity of the application itself. Today, the spotlight has shifted to the AI model and its underlying data. With techniques like fine-tuning, the same augmented dataset can be used to power multiple applications. This change has highlighted the vital importance of data in an AI-driven landscape.

The stakes are also higher now. In domains like healthcare and legal, where AI is being applied to critical decisions, high-quality data is non-negotiable. The human experts who provide feedback and label data need to be highly skilled. This underscores a shift in focus from the quantity, to the quality of data.

Why Data Quality Matters

So why exactly is high-quality data so crucial for AI?

Fine-tuning models entirely depends on the data used. Higher quality data also enables the use of smaller, more efficient models when 1) there are fewer errors in the data, 2) fewer features are needed to explain underlying patterns, and 3) overfitting to noise is less likely. This leads to improved performance, faster training times, and significant savings in compute costs. 

As a result, data scientists and engineers are becoming the backbone of any AI-powered organization. Their skills in collecting, preparing, and managing high-quality data are indispensable. Additionally, the work in data vectorization is transforming how we interact with unstructured data. By converting PDFs, images, audio, or video into embeddings, we can ask more nuanced questions and find relevant information faster. Vector databases, while still evolving, will play a central role in this new paradigm.

The Evolution of Data Pipelines

So how are data workflows and pipelines changing in response to these new realities? Here are some of the key trends we’re seeing:

1. New Data Types & Modalities: Unstructured data like text, images, audio, and video are gaining prominence. New modalities are emerging, powered by techniques like embeddings and vector search.

2. Automation & Augmentation: Data prep and analysis are becoming increasingly automated, with tools like co-pilot assistants and auto-generated code. This is freeing up data scientists to focus on higher-value tasks. An example of a company in this domain is Keebo, which automatically optimizes Snowflake parameters and queries to save engineers time and money. Another, Typedef, backed by Pear, is working to monitor DAGs within a multi-step datapipeline to develop auto-tuning and optimizations that enhance the efficiency of pipeline execution.

3. Scalable Data Infrastructure: Data infrastructure is becoming faster and more efficient to handle the demands of AI. Vector databases are enabling fast retrieval and inference on massive datasets. “RAG-as-a-Service” is emerging to connect an organization’s proprietary data with large language models. For example, EdgeDB is an open source database that enhances PostgreSQL with hierarchical queries that are more efficient in handling AI applications with tree-like structures. 

4. Collaborative Data Science: “Notebooks 2.0” — in addition to advanced platforms such as Jupyter Notebook, Databricks, and Tableau — are enabling more collaborative features and tools for accessible data science. Techniques like text-to-SQL and semantic analytics are democratizing data exploration.

5. Data Quality & Labeling: With the primacy of data quality, companies are investing heavily in data labeling and annotation. A whole ecosystem of services is emerging to provide high-quality, human-in-the-loop data labeling at scale. Synthetic data generation using AI is also being used to augment datasets. An notable example is Osmos, backed by Pear, automatically catching errors in data and removing the need for manual data cleanup.

6. Feature Stores & ML Ops: Dedicated feature stores are becoming critical to serve up-to-date features across AI models. ML Ops platforms are being adopted to manage the full lifecycle from data prep to model deployment. Versioning, metadata management, and reproducibility are key focus areas. Examples from the many existing ML Ops platforms include Vertex AI, DataRobot, and Valohai.

7. Real-time & Streaming Data: As AI is applied to real-time use cases like fraud detection and recommendations, streaming data pipelines using tools like Kafka and Flink are gaining adoption. Online machine learning is enabling models to continuously learn from new data. A prominent example is Aklivity, a company that makes a business’s real-time data streams connected to Kafka available through APIs.

8. Governance & Privacy: With AI models becoming more complex and opaque, there is a heightened focus on responsible AI governance. Tools for data lineage, bias detection, and explainable AI are being developed. Techniques like federated learning and encrypted computation are enabling privacy-preserving AI.

The Opportunity Ahead

The rapid evolution of data handling in the AI era presents a massive opportunity for data and analytics companies. Organizations will need expert guidance and cutting-edge tools to harness the full power of their data assets for AI.

From automated data prep and quality assurance, to scalable vector databases and real-time feature stores, to secure collaboration and governance frameworks — there are opportunities at every layer of the modern AI data stack. By combining deep domain expertise with AI-native architectures, data companies can position themselves for outsized impact and growth in the years ahead.

As AI continues to advance and permeate every industry, the companies that can enable high-quality, responsible, and scalable data pipelines will be the picks and shovels of this new gold rush. The future belongs to those who can tame the data beast and unleash the full potential of AI. Are you building in this space? Let’s talk.

Acknowledgements

I’d like to thank Avika Patel and Libby Meshorer for their contributions to this post. Visit our AI page to read more about the 16 spaces we’re excited about.

Navigating Security: Opportunities and Challenges in the AI Era

Posted on by lucy lee

The new generation of AI poses both huge opportunities and risks. While AI can open up a world of new capabilities, it also presents new security concerns, that require our focus at three levels:

  1. LLMs Reliability
  2. Security risks posed by GenAI
  3. AI-powered security solutions

In this article, we will explore the new reality with AI in each of these areas.

LLMs Reliability

LLMs have demonstrated remarkable capabilities in natural language processing, but their reliability remains a concern. In 2023, researchers from Stanford University discovered that GPT-4 could generate highly persuasive disinformation articles that were difficult to distinguish from real news, highlighting ongoing reliability challenges with state-of-the-art language models. We see a growing number of companies addressing issues like biased outputs, hallucinations, and the potential for generating harmful content, through improved AI infrastructure like RAG and mechanisms to test and validate LLMs.

There are a few different methods enabling to test and promise the reliability of the LLMs in our usage, among them:

1. Red teaming: Actively trying to find ways to make the model produce undesirable outputs, to identify weaknesses. Companies like Anthropic, Halcyon, and Adept AI are using red teaming in their AI development processes. Startups like Haize Labs, Robust Intelligence, and Scale AI have products helping provide solutions to handle Red Teaming.

2. Oversight sampling: Regularly sampling outputs and having them reviewed by human raters for quality and safety issues. Startups like Fiddler AI provide solutions with humans in the loop to check for quality issues

3. Runtime monitoring: Analyzing model inputs and outputs in real-time to detect potential reliability issues. Guardrails AI, Galileo and TrueEra are building infrastructure for runtime monitoring of LLMs in production.

    Security risks posed by GenAI

    Generative AI introduces new security challenges. For example, deepfakes can produce highly realistic fake content, potentially leading to misinformation and fraud, and cybercriminals are leveraging tools like Midjourney and Stable Diffusion to generate synthetic media for social engineering attacks. Additionally, GenAI systems are especially vulnerable to unique threats:

    • Prompt injection attacks attempt to craft inputs that cause the model to ignore instructions and do something else, like disclosing sensitive data. In 2023, prompt injection was used to get GPT-4 to reveal training data.
    • Jailbreaking aiming to bypassing safeguards and performing unintended actions, like creating harmful outputs or giving illegal instructions.
    • Model integrity erosion happening when an AI system’s performance deteriorates over time due to adversarial or unforeseen inputs, corrupting the effectiveness of of AI driven security measurements.

    Companies like Flow Security (now CrowdStrike), Sentra, Protect.ai and HiddenLayers are developing solutions to protect data and models from unauthorized access and malicious activity. Cohere, Anthropic, OpenAI, Adept and others are exploring new AI architectures that are more resistant to prompt attacks and jailbreaking attempts.

    AI powered security solutions

    Alongside these risks, AI offers an outstanding opportunity to address security challenges like never before. AI-driven tools can enable high-quality observability, accurate detection, clear prioritization, and accelerated mitigation. Overall, AI can transform the way we handle and mitigate security risks today. Here are a few areas with significant potential for improvement in the new era of AI:

    1. Anomaly and Threat Detection: LLMs are designed to analyze large amounts of data and identify anomalies more efficiently than humans. This enables the creation of better alert systems that detect fraud and security threats effectively and in real-time. For example, Noname uses AI to identify data leakage, suspicious behavior, and API security attacks, as they happen. Redcoat AI and Abnormal Security identify phishing attempts and malicious email activity.

    2. Penetration Testing: AI-powered tools can be used not only to test the reliability of LLMs, as demonstrated by companies like Adept and Haize Labs, but also to perform intensive and sophisticated penetration testing on systems to identify vulnerabilities, as offered by XBOW. AI-driven simulations of cyber-attacks on networks and systems can test their resilience and train cybersecurity professionals in incident handling, regularly improving security layers.

    3. Code as language: While GenAI-generated code can raise concerns among tech leaders due to potential vulnerabilities and logical flaws, LLMs can read code as if it were natural language, enabling the identification of problematic code blocks and configurations that may lead to security breaches. AI-powered tools and security-oriented LLMs like Snyk DeepCode and Codacy embody the ‘shift left’ philosophy, focusing on identifying and resolving security issues early in the development lifecycle rather than addressing them post-deployment.

    4. Vulnerability Management and prioritization: AI can be highly effective in assisting engineers with intelligent security vulnerability management and prioritization. By creating a unified source of truth for existing security vulnerabilities and analyzing factors such as severity and potential impact, platforms like Wiz and Balbix offer advanced vulnerability management and prioritization, resulting in decreased engineers confusion and response time.

    5. Incident Response and auto mitigation: AI can significantly enhance incident response and automated mitigation, like applying security patches and updates to vulnerable software components in real-time, reducing the time required to contain and resolve security breaches. Solutions like Palo Alto’s Cortex XSOAR, also leverage AI to speed up incident investigation, automate and expedite tedious, manual SOC work, towards the vision of mitigating risks with minimal human intervention.

      While the breakthroughs in AI present exciting opportunities, it is crucial to address the risks related to AI models and security. By focusing on the reliability of LLMs, understanding the new threats posed by GenAI, and leveraging AI to enhance security measures, we can navigate this new era of technology safely. Are you building in this space? Let’s talk.

      Acknowledgements: I would like to thank Pear AI Fellow Libby Meshorer for significant contributions to this post, as well as Avika Patel and Pear team members Lucy Lee Duckworth, Arash Afrakhteh, and Jill Puente for contributing.

      Human Simulating AI Agents are Closest Approach to AGI, Unlocking Value in our Everyday Lives

      Posted on by lucy lee

      AI Agents: Turning Imagination into Reality

      After sharing our GenAI thesis and the 16 fields we are particularly excited about in AI, we’re delving into one of the most interesting trends in the era of capabilities unlocked by GenAI: AI agents, and, specifically, human simulating agents.

      AI agents are software entities that can perceive their environment, make decisions, and take actions independently without human supervision. They are the closest approximation we have today to the vision of Artificial General Intelligence (AGI), replicating a broad range of human cognitive abilities, including perception, reasoning, planning, learning, and adapting to new situations without dedicated preparation.

      AI Agents Add Value Across the Board

      The AI agents space can be divided into several key subspaces and categories, some of which we have already started investing in. One agent can have multiple overlapping functionalities and several interfaces simultaneously:

      AI Agents with Specific Functionalities

      • Human-simulating agents: These agents simulate human behavior and thoughts based on a given profile or need. These transformative approaches can be applied to various use cases, from companions that fight loneliness, to agents with demographic traits, beliefs, and preferences that help predict trends like election results or consumer adoption more quickly, cheaply, and accurately. See more in our section on Human Simulating Agents below.
      • Assistant agents: These agents can handle a wide range of tasks, from running an online search, or playing a song upon voice request (e.g. Siri and Alexa), to scheduling a doctor’s appointment and maintaining a detailed to-do list, as Ohai.ai and Martin do.
      • Automation agents: These agents connect two machines, identify gaps and repetitive processes, and create automated workflows to enhance them. For example, Orby AI, backed by Pear, automates workflows for enterprises in minutes, improving team efficiency by over 60%.
      • General purpose agents: While very broad and challenging to build effectively, these agents can complete multiple tasks across different verticals and complexity levels. Key players in this space include BabyAGI, AgentGPT and Personal AI, which offer solutions for various problems through their agents.
      • Vertical agents: Highly skilled in specific fields such as healthcare, marketing, gaming, legal, and more, these agents excel in their respective areas. A promising code generation agent and engineering co-pilot is Devin, built by Cognition AI, which enables engineers to plan and execute complex engineering tasks.
      • Embodied agents: Embodied agents operate on edge devices such as IoT devices, robots, and drones, as navan.ai does. They enable smarter and more independent decision-making and actions, aiming to improve smart home systems, agricultural practices, defense, and more.
      • Collaborative agents: These agents can interact effectively with other agents, learn from each other, and cooperate to improve their performance and execution over time. Relevance AI is building such solutions to increase team productivity.

      Agents with Defined Interfaces

      • Human Facing Agents: interact with humans through text, audio, video, etc., just like Google Astra does to help people navigate their surroundings.
      • System Facing Agents: Interact with machines and systems through APIs, scripting, and data.
      • Physical world facing agents: interact with the physical world, learning and interacting with it through robots, drones, and automotive, as Tesla autopilot and Weymo.

      Agents Infrastructure

      Agents infrastructure will include ops layers that encompass memory, compute and data infrastructure. Additionally, new agent management platforms will emerge, enabling building, orchestration, observability, and monitoring. Companies like Dust and AgentOps aim to offer these solutions to help agents and their operators achieve their full potential.

      Human Stimulating Agents: Revolutionizing Customer Support, Trend Predictions, and More.

      Human simulating agents allow us to leverage AI to learn from and predict human behavior in different scenarios. There are two broad categories of Human Simulating Agents:

      Support agents

      • Assistants: As mentioned above, AI assistants can simulate human behavior while supporting us. For example, they can order and adjust our grocery lists based on our historical needs and preferences, declutter our email inboxes, and respond in our unique writing styles. They will interact with us in natural language, as if we were asking for help from a close friend, helping us manage our busy lifestyles.
      • Customer support: AI agents in the customer support space will become more sophisticated, providing relevant and satisfying support that saves man-hours, reduces customer frustration, and cuts costs for customer-facing companies. Sierra, Crescendo and yellow.ai are already on a mission to transform the customer support experience using AI agents.

      Human persona agents

      • Trend predictions: Agents will simulate people with specific demographic contexts, traits, and perspectives to predict the adoption of new consumer products, as Keplar and subconscious.ai do, create the most effective personalized marketing content, predict election results, and more. 
      • Companions: AI agents can become companions with personality and relationship history, remembering key moments in our lives, fighting loneliness, and offering initial mental health support if needed. For example, Replika and Kindroid provide engaging relationships with customizable AI companions that users can interact with through text and voice.

      Building in this space?

      If you are building in the agents space, reach out to us (Arash or Arpan). We would love to discuss your vision and explore how we can support your journey.

      Acknowledgements: I’d like to thank Pear AI Fellow Libby Meshorer for significant contributions to this post as well as Pear team members Arpan Shah, Jill Puente, and Lucy Lee Duckworth for contributing.

      Pear Biotech Bench to Business: insights on ‘Designer Immune Systems,’ allogeneic stem cell therapies, and making an impact on the lives of patients with Ivan Dimov

      Posted on by jill

      Here at Pear, we specialize in backing companies at the pre-seed and seed stages, and we work closely with our founders to bring their breakthrough ideas, technologies, and businesses from 0 to 1. Because we are passionate about the journey from bench to business, we created this series to share stories from leaders in biotech and academia and to highlight the real-world impact of emerging life sciences research and technologies. This post was written by Pear PhD Fellow Sarah Jones.

      Today, we’re excited to share insights from our discussion with Dr. Ivan Dimov, CEO and co-founder of Orca Bio. Ivan has co-founded three high-tech companies and two R&D centers, and he is now working to make next-generation cell therapies safer and more efficacious at Orca.

      More about Ivan:

      Ivan earned a Ph.D. in Applied Biophysics from Dublin City University and has since worked as a postdoc at UC Berkeley and as a visiting instructor and senior scientist at Stanford. His passion for translating his work and his tech-heavy background have made him an expert in electronics and bio-microelectromechanical systems (bio-MEMS) and have helped him to work on numerous projects and companies including Blobcode Technologies, Lucira Health, and Orca Bio. 

      If you prefer listening, here’s a link to the recording! 

      Insight #1: Instead of building a technology and then searching for the right application, it’s much more efficient to identify the right problem prior to creating a solution. 

      • There are many approaches to starting a company and making new, impactful discoveries. As someone with a strong tech and engineering background, Ivan was trained to find and create interesting, powerful new technologies and go hunting for applications; he made hammers and went searching for nails. 
      • However, when working with Dr. Irv Weissman at Stanford as a postdoc, Ivan learned to do things a little differently. 
      • Though Ivan’s background was primarily in applied biophysics and bioengineering, the Weissman lab’s focus was medicine and biomedical research. Ivan acknowledged that when you work in medicine, you are exposed to an over-abundance of problems, and in this environment, Ivan learned that the most effective solutions are those that are tailor-made to fulfill a clearly defined need. 

      Working with physicians was a huge change in mentality for me…you’re seeing suffering everywhere and you have all these problems, and you sort of have to figure out, okay, which problem do you want to focus on, and what’s the best solution or technology you can come up with for that problem? I think that’s probably the better way of doing innovation…rather than trying to squeeze in some technology that was thought of in a different context and trying to make it work.

      • One such problem was related to the poor outcomes of stem cell transplants, a procedure in which a donor’s stem cells are harvested and administered to a recipient. Ivan explained that there wasn’t a way to sort out the good cells from the bad and ensure that the recipient was only receiving cells that would therapeutically benefit them and minimize unwanted side effects. 
      • The idea of creating precise and well-defined stem cell therapies would become the central theme of Ivan’s work at Stanford and later of Orca Bio.

      Insight #2: Academia is great for exploring, learning, and making mistakes. However, industry is where you can iron out the more mundane details of company creation and focus on impact and real-world use cases.  

      • Having started three companies–Blobcode Technologies, Lucira Health, and Orca Bio–Ivan has extensive experience in taking ideas from academia to industry. 
      • Lucira Health, a diagnostics company that has since been acquired by Pfizer, was spun out of Ivan’s work at Berkeley. The goal was to miniaturize a microfluidic chip that could be utilized as an at-home diagnostic. Notably, the company received approval from the FDA for their at-home COVID test that could read out results in about 30 minutes.
      • While at Berkeley, Ivan spent time fine-tuning the idea and conducting proof-of-concept experiments for the chip. However, it became apparent that this academic setting wasn’t necessarily conducive to the less thrilling aspects of the project. Spinning out and starting Lucira allowed the team to more efficiently work on the ‘mundane’ details like reproducibility and clinical trial design.

      [Academia is a safe place where] there’s a lot of openness to trying out new things… and the greatest thing about it is that you can try it and you can make a mistake and that’s okay. You can come up with a better alternative.

      • While Ivan agrees that academic labs are a great place for ideation and company incubation, it’s important to be vigilant and humble enough to realize when it’s time to take the next step. Industry and academia each have their respective strengths, and Ivan learned that both were crucial to the growth and future success of his companies.

      Insight #3: Stem cell therapies don’t have to be so risky: by cherry-picking the cells that a patient receives, long-term outcomes can be significantly improved.

      • In leukemia, cells in the bone marrow and lymphatic system become cancerous and rapidly multiply. To treat this type of malignancy, patients often go through multiple rounds of chemotherapy, radiation, or targeted immunotherapy and may receive an allogeneic stem cell transplant. 
      • Essentially, a conventional allogeneic transplant begins when the patient receives chemotherapy and/or radiation to wipe out all of the cancerous blood cells together with the patient’s healthy blood and immune cells. Once the cancer can no longer be detected, stem cells from the bone marrow of a healthy donor will be administered. These new cells can multiply and grow into mature, functioning blood and immune cells.
      • For some patients, this treatment is curative and wipes out any trace of cancer from their systems. However, even after chemotherapy and radiation, some cancer cells may go undetected and cause a patient to relapse.

      The problem with cancer is that if you leave even a little bit of it behind, even a single cell that hides and survives, it has the potential to reinitiate and restart your cancer from scratch… When you get into full remission–meaning we can’t measure any more cancer in you–it just means that our tests aren’t sensitive enough to see if it’s there or not there.

      • Once the stem cell transplant is complete, it takes a couple of weeks for the new immune and blood systems to get up and running. The hope is that these new immune cells can wipe out any remaining cancer cells that may be hiding out. 
      • In addition to the potential for relapse, patients frequently develop either acute or chronic Graft-vs-host-disease (GVHD), complications in which the new immune cells from the donor start to attack the patient’s (host’s) own cells and tissues. GVHD can affect many parts of the body and can even lead to death. 

      In a standard transplant, your chances of surviving for twelve months free of relapse or free from GVHD is somewhere around 30-40%. With Orca Bio… we can get rates somewhere between 70-80% ideal survival rates.

      • So how do they do it? Ivan’s goal at Orca Bio is to revolutionize the cell therapy space by creating a high-precision cell therapy that gives patients only the most efficacious donor cells. 
      • Orca’s unique platform identifies and sorts for donor cells that have the highest therapeutic benefit. By removing cells that either harm or don’t help the patient, the patient’s chances for relapse or developing GVHD are dramatically reduced.
      • With what they call their ‘designer immune system,’ Orca’s approach aims to help patients recover more quickly, prevent relapse, and be safe enough for older or sicker patients who can’t receive traditional stem cell transplants.

      Insight #4: To solve the problems of current allogeneic stem cell transplants, you have to balance killing the remaining cancer cells with protecting the patient’s own tissues and cells.

      • When designing an immune system to infuse into patients with blood cancer, it can be difficult to kill cancer cells without harming other cells in the patient’s body. 
      • In a healthy immune system, cells called regulatory T cells (T-regs) monitor and regulate what effector T cells are doing. Such effector T cells can help promote inflammation and eliminate cancer cells. However, when a patient has cancer, there is an imbalance between these two cell types, and the immune cells don’t effectively kill the cancer cells. 
      • These types of cells are often involved in autoimmune disorders and can also play a role in the development of acute GVHD shortly following the stem cell transplant or in chronic GVHD, long after the treatment has concluded. 
      • Orca Bio’s first product, Orca-T, helps to restore balance in the immune system by first bringing stem cells and T-reg cells into the patient’s body to let them set up the immunoregulatory environment. Once the T-regs and stem cells have had a chance to settle in and begin restoring the patient’s immune and blood systems, conventional T cells with cancer-killing capabilities are administered.
      • Orca-T has reached Phase III clinical trials for indications such as acute myeloid leukemia (AML), myelodysplastic syndrome (MDS), acute lymphoid leukemia (ALL), and mixed-phenotype acute leukemia (MPAL) in patients with matched donors who are younger than 65. Matched donors are those that share the same human leukocyte antigen (HLA) profile, and this means that these cells are less likely to be identified as intruders in the patient’s body, thus reducing the risk for GVHD.
      • Patients receiving Orca-T first receive chemotherapy and/or radiation to target cancer cells and suppress their immune systems. The first dose of Orca-T is an infusion of stem cells that regenerate the blood and T-regs that help set the immune landscape. Two days later, the patient receives an infusion of conventional T cells that can begin to attack any remaining cancer cells. 

      What’s amazing about this approach is that by doing that, you’re not turning off the effector T cells from destroying the cancer. You’re just turning off alloreactivity in the key organ sites where you might create GVHD, but you’re still keeping it on for wherever the cancer might be.

      • Moving forward, the company is continuing its work on Orca-T by expanding the age range of patients who can be treated with the drug.
      • The pipeline also includes next-gen cell therapy treatment, Orca-Q.
      • To solve the problem of limited matched donor availability, Orca-Q is a high-precision cell therapy that has been tailored for haploidentical, or half-matched donors. These donors are typically parents, children, or siblings and can be much easier to find. However, the risk for GVHD increases with a half-matched donor compared to a fully matched donor.
      • Orca-Q has so far shown positive results in Phase I in oncological indications and is being investigated for autoimmune and hematological indications, as well. 

      Insight #5: Sometimes science is personal: reflecting on Orca’s journey, Ivan and his team have a deep understanding of how their work can change lives.

      • Having treated more than 400 patients so far, Orca has seen firsthand how patients can benefit from their novel stem cell transplants. 
      • In particular, patients who are too old or too sick for traditional transplants now have a fighting chance.

      One of the most incredible stories was about my co-founder’s [Nate Fernhoff’s] father-in-law. He was 71 years old when he was diagnosed with myelodysplastic syndrome. He had an aggressive variant of the disease…however, physicians feel very skittish about treating folks at that age with a myeloablative allogeneic bone marrow transplant, so they’re offering a reduced protocol [with a much worse chance of controlling the cancer]. We started looking for clinical trials, anything that would cover folks of that age.

      • Dr. Fernhoff’s father–in-law, Mikhail Rubin, was diagnosed with a rare form of blood cancer and found that his options for treatment were extremely limited. The most successful and aggressive forms of treatment were offered only to younger patients. 
      • Meanwhile, Orca’s clinical trials had so far proven to be safe and effective. The Orca team sprang into action and started working to convince physicians and the FDA to allow them to treat patients older than 65 and expand the enrollment criteria for the trial.
      • Ivan noted that this exclusion of the most dire patients stems from the industry’s hesitancy to add further risk to clinical trials.

      In April of 2021, we were given the permissions, and were able to treat him. It’s been a phenomenal recovery. He recovered much faster than any of his younger counterparts even though a lot of physicians thought it would take months in the hospital for him to get out. Yet, in the first year after, he started riding his mountain bike and did 3,000 miles on his bike.

      • Not only is Mr. Rubin back to biking, he has also been cancer-free and GVHD-free for three years now. 
      • While science tends to be objective in nature, personal connections and motivations help drive the mission and make work like this possible. 

      PearX W25 applications are now open

      Posted on by jill

      The S24 batch of PearX is underway and the founders have been pushing hard to launch products, grow customers, and get their products to scale. This year we have seen a surge in AI companies with over 95% of S24 companies focusing on AI. Our AI team is ready and looking for companies in AI infrastructure and vertical applications. Read more about our AI thesis. We couldn’t be more excited about the S24 batch of companies, which we will share more about publicly in the fall.

      Simultaneously, we have been hard at work preparing for the W25 batch kicking off in January. Today, we’re opening applications for our W25 cohort and encourage all teams from idea stage to companies with traction to apply. Apply here now.

      Key Dates:

      Early applications and interviews: Early application deadline on August 30th. If you apply by this deadline, you will hear from us by mid-September.

      Regular Applications: Applications close on October 1st. You will hear from us by November 25th at the latest.

      PearX is an immersive small-batch 14-week accelerator which counts top performing companies like Viz.ai, Affinity, Xilis, Capella Space, Cardless, Nova Credit, Federato, Valar Labs among its alumni. Pear has also backed DoorDash, Dropbox, Vanta, Aurora Solar, Gusto, Guardant Health at the seed stage.

      At PearX, we provide custom services and support for each company. From capital, hiring, founder-led sales, to fundraising support, everything we do takes into account the unique needs that each company has.

      Here is what you get with PearX:

      • Capital to build your company, your way: We invest between $250k and $2M in all PearX companies. We know that some founders only need a small amount of capital to ideate and other teams are in more cost-intensive verticals that require more funding. All companies are unique so we’ll work with you. 
      • Credits: We provide up to $650k in credits from top providers like Azure, OpenAI, AWS, Google, Anthropic, and many more.
      • Founder to founder: Work 1:1 with a partner who has been in your shoes and knows your industry. Our team has started and sold 10 companies to the likes of Cisco, Instacart, Plaid, and Zynga.
      • Join the best community: Entrepreneurship doesn’t have to be a lonely journey. Joining PearX means joining a community of like-minded founders. We kick off each cohort with Camp Pear: a 3-day retreat for the entire cohort to come together, learn key company building tactics, and get to know one another. Not only will you work alongside your PearX batch for 14+ weeks, but you’ll also have the wider PearX alumni network to lean on. 
      • Access Pear Studio: Everyone in PearX receives dedicated office space in Pear Studio SF, our 30,000 square foot state-of-the-art office space with standing desks, conference rooms, phone booths, and more. This space is completely free to you for the first 12 months.
      • Build a scalable sales motion: Our go-to-market team, Pepe and Ana, will guide you through the critical steps of sales: nailing ICP, prospecting, customer discovery, messaging, and more. 
      • Recruit the best talent: Our dedicated PearX Recruiter, Nate, will find your founding engineer, co-founder, or whatever critical hire your team needs. In the last two cohorts, Nate has hired 25 people for our PearX companies. Nate leads the full cycle of recruiting for your team – from sourcing to closing candidates. This is an unprecedented level of support for an accelerator, but that’s how much we believe that hiring impacts company building. Last batch, Nate hired an average of two people per PearX company.
      • Fundraise strategically: We help you raise additional capital when you’re ready. From perfecting the story and creating a pitch deck to creating a target investor list and negotiating and closing your round. In fact, 90% of companies that go through PearX raise capital from institutional investors.

      Join us for PearX W25:

      Are you interested in joining our PearX W25 cohort? We’re looking for the next generation of category defining companies. Please apply at pear.vc/pearx.

      How three PearX S19 alums raised $50M this quarter

      Posted on by jill

      Today, raising capital is far more challenging than it was a few years ago. There are no fake Series A’s— most businesses require a good foundation, strong unit economics and growth as well as a moat to get there. We talked to three alums from our 2019 PearX cohort to hear how they raised capital this year: Andrew Powell from Learn to Win, John Dean from Windborne, and Parth Shah from Polimorphic. 

      PearX is our exclusive, small batch, 14 week program. 90% of our companies go on to raise a successful seed round from top tier investors. PearX alumni companies include Affinity (S14), Viz.ai (S16),  Cardless (S19), Federato (S20), Valar Labs (S21), and more. 

      Answers have been edited for brevity. 

      Learn to Win (PearX S19):

      Learn to Win is a training platform that empowers companies to design, deliver and assess the impact of employee training. They serve customers across commercial and government markets in primarily high-intensity training situations. Learn to Win closed a $30M Series A round in June 2024, led by the Westly Group and joined by Pear and Norwest Venture Partners.

      Windborne Systems (PearX S19):

      Windborne is a full-stack, vertically-integrated weather intelligence company. They operate the largest balloon constellation on the planet, running a base weather forecast with the data they collect from those balloons. Windborne raised a $15M Series A round, led by Khosla Ventures and joined by existing investors Footwork VC, Pear VC, and Convective Capital.

      Polimorphic (PearX S19):

      Polimorphic digitizes government operations, helping governments provide a great customer service experience to their residents and businesses. AI-powered search and voice software empowers municipal employees with saved time and resources, while delivering a modern service experience that delights residents. Polimorphic raised a $5.6M round, led by M13 with participation from existing investors Shine Capital and Pear VC.

      It’s been five years since you went through PearX. What did you learn from this accelerator program that still provides you value today? 

      Windborne: Mar played a big role in us even founding a company. Our very first check came from Pear Dorm. Fundamentally, we learned how to be entrepreneurs, and every connection in those early days came through Pear. 

      Polimorphic: One of the most interesting pieces of it was thinking about how venture-scale businesses are different. When we came into PearX, we didn’t even have a company yet, and we hadn’t even landed on this iteration of Polimorphic. 

      It takes time to find product market fit. PearX supported us during the exploration phase— you need to be nimble, trying stuff until you find what clicks. COVID interfered with a lot of our plans to work with the government in the early days, and recently we’ve really found product market fit.  

      Where has Pear helped your company the most? 

      Learn to Win: Hiring and fundraising. Pear has helped hire our first few engineers and our first few executives. 

      Windborne: Especially in the last year and a half, Pear’s talent team has been insanely helpful with hiring. Beyond that, they offer general advice on people ops: compensation, policies, equity splits— everything around managing people. We use Ashby for free through Pear. It’s very convenient to have a VC you trust a lot to help you get set up with these things. 

      One of the single biggest challenges that all companies face is talent. If you’re good at talent, you’ll win. If you’re bad at talent, you’ll lose. To have a VC that stands out in talent is incredibly valuable. 

      What’s a piece of advice for founders trying to raise a Series A? 

      Learn to Win: When it comes to Series A, there’s certainly more of an emphasis on metrics and scalability. Talk to other folks in your sector and ask them what key metric they were gunning for, what pieces of evidence they gave to investors to help them understand your business opportunity.

      At seed, we proved we could build a great product and deliver great value for customers. What needs to change is an engine that can repeatedly do that at scale. 

      Windborne: My biggest piece of advice is to be wary of advice that isn’t relevant to you. Think about where you want advice and where you don’t. For us, we disregard a lot of things people tell us when it comes to engineering and manufacturing when they don’t understand the way we run our business. You’re not going to win by only following commercial industry wisdom. 

      As the rounds progress, things definitely get harder— riskier, scarier, more challenging. But they also get so much more fun. 

      What sets Pear apart from other venture firms? 

      Polimorphic: Pear’s willingness to be involved. We were pre-idea, pre-company when we met Pear; we just knew we were interested in the political realm. Pear stuck with us as we started to explore government. Their ethos of investing in people manifested in sticking with us through a few different iterations.

      Learn to Win: Pear is an expert at early stage— even what comes before it. We were still students at Stanford working out of our dorm room, and Pejman and Mar were some of the first people to believe in the potential of our business. In the early days, there’s so many people that could point out a million reasons why your startup will fail. Pear believed in the one reason out of a million and helped us understand how to dig into it. 

      What are your goals for the next phase of your company? 

      Learn to Win: We’re trying to ramp up our defense business and grow aggressively across the board. We’re just scratching the surface, and we’re excited to see what we can do with this new capital.

      Windborne: We want to scale up data collection operations. We’re launching around 100 balloons a month, and we want to be doing a few hundred a month. By the next round, we want to be doing over 10 million a year in revenue. And more importantly, we want to be collecting more in situ weather observations than the rest of the world combined.

      Polimorphic: We’re on path to having 100 government clients, which is a big milestone. We started with cities and counties and are about to do our first state-level deal. That’s the next big phase: acquiring more customers, delivering a new paradigm for governments to provide customer service to their residents. 

      All three of these companies were in the same PearX S19 cohort. At the time, these startups were all just a few founders and an idea. It’s amazing to see their growth and to see them collectively raise $50M+ over the last quarter. Read more about PearX here. Applications for PearX W25 open on August 20th.

      Camp Pear: PearX S24

      Posted on by jill

      PearX is our accelerator for early stage companies and our S24 batch is officially underway. We welcomed 22 companies into PearX’s S24 class at Camp Pear: a 3-day immersive retreat for the entire cohort to come together, learn key company building tactics, and get to know one another. This class of PearX founders is overwhelmingly comprised of AI companies, with 21 of the 22 being AI-first. As is now our tradition, we kicked off the cohort with Camp Pear. 

      Practical advice from PearX alums

      The founders learned from expert speakers about key company building tactics – from hiring to go-to-market to fundraising and more. And they heard from Pear founders who have been in their shoes before – like Ray Zhou, Co-founder of Affinity (PearX S14), Scott Kazmierowicz, Co-founder and CEO of Cardless (PearX S19), Nicole Rojas, Co-founder of Lasso (PearX S22) and Lucia Huang, Co-founder and CEO of Osmind.  

      The power of community

      Camp Pear provides a unique way for founders to build community right from the beginning of each cohort. Through team building activities, bonfire conversations, and shared meals – founders got to know each other on a personal and professional level. Building a company from scratch is really hard, and we believe community plays a key role in the success of an entire cohort. 

      Kickoff to a fast 14-week company building sprint

      PearX starts at Camp Pear, but it’s really just the energizing start to a 14-week program. The 14 weeks between Camp Pear and Demo Day are all about company building. During the cohort, each startup will work closely with the Pear team to get their startup off the ground. This includes weekly 1:1 meetings with their primary Pear Partners, weekly talks from experts across the tech and startup world, hiring help from Nate, founder-led sales training from Ana, fundraising bootcamp from Mar, and loads of fun activities (like our upcoming founder Olympics competition!). It all culminates in Demo Day, where the PearX founders will pitch to thousands of top tier investors – and each batch, 90%+ of our companies go on to raise capital. 

      We’re so incredibly excited for this S24 cohort of PearX. If you’re interested in learning more about Pear, applications for our next cohort will open August 20th for W25. Stay tuned!