Emergence Capital

Accessing Corporate Data: The Key to Unlocking AI’s Next Innovation Cycle The Emergence Capital team has made a compelling argument in their recent piece on applying innovation cycles to the latest AI platform breakthroughs. They contend that accessing corporate data is crucial for unlocking AI’s next innovation cycle. Historically, technological advancements often follow an S-Curve pattern, characterized by rapid innovation followed by stabilization. This pattern is evident in the development of networking protocols in the 60s-80s, browser technology in the 90s, and more recently, mobile application development. Applying the S-Curve concept to Large Language Models (LLMs), we have witnessed multiple breakthroughs with the releases of ChatGPT in 2022 and GPT-4 in March last year. However, recent advancements have been incremental in nature. Two primary reasons contribute to this AI ‘plateau’: 1. Exhaustion of Public Data Sources: The latest versions of foundation models were trained on the entire corpus of the internet, leaving public data sources depleted. 2. Suspect Quality of Public Data: Public data sources are often of questionable quality, particularly when large training runs indiscriminately ingest niche subreddits and forums. In contrast, data produced in a work context is more valuable and abundant than public data. Imagine foundational model companies gaining unfettered access to the data trapped in work deliverables, corporate databases, and enterprise software solutions. So, what are the opportunistic use cases to pursue? According to the Emergence team, four key areas are emerging: 1. Engaging Experts: Source high-quality data from experts in each field using novel incentive structures and community engagement. Examples include Centaur Labs and Datacurve (YC W24). 2. Leveraging Latent Data: Help enterprises prepare and utilize data from business apps for AI training. Examples include unstructured.io and Shelf. 3. Capturing in Context: Develop methods to capture new data without disrupting workflows, including multimodal content. Examples include Textio and Superlinked. 4. Securing the Secret Sauce: Assist enterprises in creating and deploying their own custom models to protect proprietary IP. Examples include Together AI and Holistic AI. ----- 🔗 For the full investment thesis, check out ‘The AI Plateau Is Real — How We Jump To The Next Breakthrough’ from Emergence Capital. 🔗 Find "Read Capital Chronicles 24" on my profile for original materials, additional takeaways, and more. #Technology #Entrepreneurship #VentureCapital

🤔 If you’re wondering how Arcee AI is consistently releasing Small Language Models (SLMs) whose performance rivals #LLMs – there’s no secret to our success. ✨ It’s because our technology is powered by a world-class research team that’s constantly updating our model training pipeline, making it possible to get MORE out of models that are increasingly SMALL. 🎉 We’re proud to announce the latest publication by our researchers, "𝗠𝗲𝗿𝗴𝗶𝗻𝗴 𝗶𝗻 𝗮 𝗕𝗼𝘁𝘁𝗹𝗲: 𝗗𝗶𝗳𝗳𝗲𝗿𝗲𝗻𝘁𝗶𝗮𝗯𝗹𝗲 𝗔𝗱𝗮𝗽𝘁𝗶𝘃𝗲 𝗠𝗲𝗿𝗴𝗶𝗻𝗴 (𝗗𝗔𝗠) 𝗮𝗻𝗱 𝘁𝗵𝗲 𝗣𝗮𝘁𝗵 𝗳𝗿𝗼𝗺 𝗔𝘃𝗲𝗿𝗮𝗴𝗶𝗻𝗴 𝘁𝗼 𝗔𝘂𝘁𝗼𝗺𝗮𝘁𝗶𝗼𝗻." The paper introduces DAM as an efficient, data-informed, adaptive merging approach that can be an alternative to evolutionary merging. DAM optimizes model integration through scaling coefficients, which minimizes computational demands. The paper also explores model merging techniques across a spectrum of complexity, examining where automated methods like evolutionary strategies stand – compared to hyperparameter-driven approaches such as DARE and TIES-Merging, and and also compared to simpler methods like Model Soups. Our findings challenge the traditional assumption that more complex methods are inherently superior, showing that straightforward techniques like linear averaging can perform just as well, especially when merged models share similar characteristics. Huge congrats to our researchers Thomas Gauthier-Caron, Shamane Siri, PhD, Elliot Stein, Malikeh Ehghaghi, Charles Goddard, Mark McQuade, and Jacob Solawetz... and with a special shout-out to our longtime collaborator, co-author Maxime Labonne. Read the paper here ⬇ https://lnkd.in/e_DdYBDm #NLP #GenAI

I am often asked by LPs who is the greatest seed investor of our time. The answer is simple: David Frankel Founder Collective ------------------------ Coupang: $46BN market cap SeatGeek: $1BN valuation PillPack: $1BN acquisition Suno: $500M valuation ------------------------ Total: $48.5BN value created ------------------------ My top 7 lessons 👇 1. Has DPI Died in 2024 - Our fund III (2018 vintage) still has no DPI. - Even our smaller hits in Fund I and II have given DPI. - It was a perfect storm of ZIRP & day trading during COVID. 2. Reserves Is the Toughest Aspect of Venture - When we reserve more, I’m worried the market will go crazy. - When we reserve less, I’m worried the companies will need more funding in the future. - It is hard to get it perfect. 3. The Hard Truth About Seed Fund Economics - There is more polarisation than ever before. - You need to have 4-5 good fund returners. - It’s better to sell 4 companies at $250M each than 10 companies at $100M each. 4. Never Sell Your Winners - Uber, Trade Desk & Coupang all returned the fund - I would have never sold a single thing. - You want to be in your winners forever. 5. Casino Capitalist and the Sin of the Era - In 2 companies I invested in, John Curtius @ Tiger wrote a $20M on $80M check after one call. - The boatload of capital was the biggest sin of the last era. 6. How Founders Can Know if an Investment Is Meaningful to a VC - They have to join the board. - A $5M check with a board seat is a better signal than $10M with no board seat. - It shows they are willing to spend time there. 7. Why Failed Founders Tend to Be More Successful Than Super Successful Ones  - Super successful founders come back & think they can conquer anything. - Generally that has not worked out. - Failed founders come back hungrier: They have a chip on their shoulder.

Yet another Arcee AI Small Language Model (SLM) that's punching above its weight. 💪 Our 14B SuperNova-Medius is #1 on the Hugging Face Open LLM Leaderboard for models under 21B parameters!! We released it on Friday and it's already been downloaded nearly 8k times. Check it out here: https://lnkd.in/ega2NEH6. And learn more about how we trained SuperNova-Medius using cutting-edge techniques that we've pioneered, including distillation and model merging, in our blog: https://lnkd.in/gDxf5m9T #NLP #GenAI #LLMs

💫 LOGIK LEGENDS 💫 What. A. Week. We've arrived at Friday, wrapping up a hugely exciting week here at Logik.io. And of course, Friday means we get to highlight someone who has played an important role in Logik.io's growth journey. Today, we're thrilled to feature... Godard Abel! As one of the "godfathers of CPQ" alongside our CEO Christopher Shutts, Godard's insight and expertise has been woven into the fabric of Logik.io since the very beginning. He has been integral in the growth of our company, serving as Executive Chairman on our Board and seemingly always making time to support our team and celebrate our successes. As CEO of G2, Godard is known far and wide as an incredible leader who is always pushing limits and amplifying growth with innovation and speed. Colleagues have described Godard as "inspiring," "approachable," "engaging," "a champion for the customer," "a caring mentor," and the list goes on... Godard truly has an amazing talent for scaling companies while always prioritizing the people around him. One teammate said: "𝑷𝒆𝒐𝒑𝒍𝒆 𝒍𝒐𝒗𝒆 𝒘𝒐𝒓𝒌𝒊𝒏𝒈 𝒖𝒏𝒅𝒆𝒓 𝑮𝒐𝒅𝒂𝒓𝒅'𝒔 𝒍𝒆𝒂𝒅𝒆𝒓𝒔𝒉𝒊𝒑 𝒃𝒆𝒄𝒂𝒖𝒔𝒆 𝒉𝒆'𝒔 𝒂𝒍𝒘𝒂𝒚𝒔 𝒑𝒖𝒔𝒉𝒊𝒏𝒈 𝒊𝒏𝒏𝒐𝒗𝒂𝒕𝒊𝒐𝒏 𝒊𝒏 𝒂 𝒘𝒂𝒚 𝒕𝒉𝒂𝒕 𝒊𝒔 𝒆𝒙𝒄𝒊𝒕𝒊𝒏𝒈 𝒂𝒏𝒅 𝒊𝒏𝒔𝒑𝒊𝒓𝒊𝒏𝒈. 𝑯𝒆 𝒑𝒖𝒔𝒉𝒆𝒔 𝒉𝒊𝒔 𝒕𝒆𝒂𝒎𝒔 𝒕𝒐 𝒕𝒉𝒊𝒏𝒌 𝒃𝒊𝒈𝒈𝒆𝒓 𝒂𝒏𝒅 𝒊𝒔 𝒔𝒐 𝒄𝒐𝒎𝒎𝒊𝒕𝒕𝒆𝒅 𝒕𝒐 𝒊𝒏𝒗𝒆𝒔𝒕𝒊𝒏𝒈 𝒊𝒏 𝒑𝒆𝒐𝒑𝒍𝒆 𝒂𝒏𝒅 𝒉𝒆𝒍𝒑𝒊𝒏𝒈 𝒕𝒉𝒆𝒎 𝒓𝒆𝒂𝒄𝒉 𝒕𝒉𝒆𝒊𝒓 𝒈𝒐𝒂𝒍𝒔." Godard, we can't thank you enough for all the encouragement, mentorship, ideas, and investment that you've provided to Logik.io over the last few years. You are the epitome of a legendary leader! Want to help us celebrate Godard? Let's hear it in the comments ⤵️

It's been 48 hours since we announced Unify's $12M Series A and reception has been WILD So far we’ve seen: 1) $2 million of pipeline created 2) 500,000+ social impressions and thousands of engagements 3) 7 articles covering Unify We spent 3 months preparing for this announcement. Here’s everything that went into it: 1) Unify 2.0 - Introducing Plays & AI Agents Our Linear initiative for the fundraise announcement had 12 projects. Our product team spent many late nights thinking, planning and building Plays and AI Agents which are central to Unify 2.0. We released 2.0 at 12:30am and at 12:42am, one of our customers messaged us “This agents feature is 🔥”. There’s a link to our Product Hunt in the comments. 2) A New unifygtm.com We re-worked our whole website, and introduced new product pages, a high-quality demo video, and a new self-service product demo. Our new website reflects feedback from customers and prospects that they wanted more visibility into our product. 3) Linkedin & X Announcements A fundraise announcement is a unique opportunity to create an echo chamber. Our whole team used this as an opportunity to share personal stories about why we’re excited to be building Unify. That plus organic shares from customers, investors, and our network amplified the effect. 4) Press Coverage from TechCrunch, Axios and more Our Series A announcement was picked up in 7 total outlets, including TechCrunch and Axios. We’re grateful to be able to tell our story in these places. 5) Unify Happy Meals On Tuesday night our SF and NYC teams spent hours packing Unify-branded happy meal boxes full of aprons and warm cookies. On Wednesday AM we shipped these out to prospects, customers and investors. Our goal was to delight people with an outside-the-box experience. And the reaction was *warm* (see the photo below)

The good news keeps coming... Following the announcement of our $25M Series B earlier this week, Logik.io has been named a Top 10 Startup in Chicago. This recognition from LinkedIn is a testament to our strong growth and our proven ability to attract the best talent in the business. I've said it before: at Logik.io, we have the best platform and the best people that are enabling us to change the CPQ technology landscape. 🔗 https://lnkd.in/gVBfBGHi We are very proud of our roots here in the Chicago area, and I want to extend my congratulations to the other startups in the Top 10 list as well. Thank you to LinkedIn for this honor, and many thanks to the entire Logik.io team for your hard work and energy that make this all possible.