Continuous Learning_Startup & Investment

Earlier today, Cisco announced its intention to acquire Splunk for $28b , a 30% premium to the closing price. Reviewing the financials, we see Splunk is a very healthy business. $3.6b in revenues growing at 37% places the business in the top quartile of public software companies. The 78% gross margin is 6 percentage points greater than the public median. The estimated sales efficiency at 0.6 is top quartile. This means for every sales and marketing dollar invested the company purchases an additional 60 cents of gross profit in the next period. The net income margin of negative 7.6% is in line with most other software companies. Now let's examine the valuation of the business. before the announcement of the acquisition a company traded at a 4.2x forward revenue multiple. The implied multiple after the M&A is 5.7x, assuming the acquisition price is at $28b. Our internal model predicts the forward multiple based on similar companies should be about 7.5x. It was a bit overoptimistic with the Klaviyo valuation (predicted $10b ; it opened at $9.3b & actual as of today is $8.5b) but this may be due to the recent downdraft in the market after the Fed indicated a longer period of high-rates & alluded to potentially raising rates once more this year. The Cisco acquisition continues the trend of take-privates that includes New Relic for $6.5b, Qualtrics for $12.5b, & Software AG for $2.4b - all in the last 6 months.

How Jeff Bezos produces unique insights: “Jeff has an uncanny ability to read a narrative and consistently arrive at insights that no one else did, even though we were all reading the same narrative. After one meeting, I asked him how he was able to do that. He responded with a simple and useful tip that I have not forgotten: He assumes each sentence he reads is wrong until he can prove otherwise. He's challenging the content of the sentence, not the motive of the writer. Jeff, by the way, was usually among the last to finish reading.”

I sent 53 cold emails to Marc Benioff, and each was individually crafted with different hooks, different personal anecdotes, I A/B tested cold emails to the extreme and finally, it worked. From the future of SF, the balance of in-office vs work from home, and the future of AI, this is a truly special one. Coming on Monday! #founder #funding #business #investing #vc #venturecapital #entrepreneur #startup #salesforce

Most people missed one of the most important news of the summer - open-access multi-modal (image + text) model coming out of Hugging Face! You can optimize it, fine-tune it and customize it for your use-case: https://lnkd.in/eWCGJ3Fn You can try the fun version of it here: https://lnkd.in/eMQA66bq or the general playground here: https://lnkd.in/epbMhXxX What will/should the community build with it?

Is only summarization dead? Or will new type of ai replace traditional computings?

중국의 Summarization 연구하던 연구자들이 화가 많이 났네요. 이제 Summarization 연구는 더 이상 끝났다고 논문을 통해 선언합니다. Summarization is (Almost) Dead https://arxiv.org/abs/2309.09558 GPT와 같은 Large Language Model (LLM) 들이 너무나 좋은 성능을 보여주면서, 기존의 많은 자연어 처리 연구 주제들을 무용지물로 바꾸어 버렸습니다. Summarization은 대표적인 자연어 처리 연구 주제 중에 하나이고, 이를 개선하기 위해 정말 다양한 연구들이 활발히 진행됐었구요. 그런데 이제 그 어떤 방법을 써도 LLM들이 할 수 있는 것에 못 미치는 상황이 되었습니다. LLM 연구자들은 보통 summarization은 일종의 test set일 뿐 실제 이를 개선하기 위한 많은 노력을 하지도 않았는데요. 먼가 산업 혁명 시대 변화와 오버랩되네요.

Introducing Sequoia’s latest market map! Health is so fundamental to our quality of life, but healthcare is rife with administrative inefficiencies. Healthcare expenditures are almost 20% of U.S. GDP. Disruption in this space has always been difficult, but gen AI is poised to change the healthcare industry – not just driving efficiency, but dramatically improving patient outcomes. We are excited to meet the innovators driving change in this space. Please give us feedback and let us know what we missed! https://lnkd.in/gmcEQ4d4

젠신 임팩트를 개발한 miHoYo는 최신 AI 에이전트에 대한 86페이지 분량의 설문조사 보고서를 공동 집필했습니다. 참고로 겐신 임팩트는 연간 10억 달러 이상의 매출을 올리는 역대 최대 규모의 오픈월드 모바일 게임 중 하나입니다. 이 논문은 인지 핵심(추론, 기억, 계획), 지각(텍스트, 시각, 청각), 행동(텍스트, 도구 사용, 구현), 참여자(단일 에이전트, 멀티 에이전트, 인간-AI 혼합), 환경(가상 샌드박스, 물리적) 등 여러 축을 고려한 훌륭한 개요를 제공합니다. 스탠포드 스몰빌('제너레이티브 에이전트')을 겐신이나 다른 대형 MMORPG의 플레이어 기반까지 확장할 수 있다면 꽤나 멋진 광경이 될 것입니다. 인간과 AI 에이전트 모두 제가 예측할 수 없는 놀라운 돌발 행동을 할 수 있을 테니까요. 안타깝게도 효율성과 비용이 이러한 대규모 배포를 가로막는 주요 과제입니다. 일단은요. 논문 "대규모 언어 모델 기반 에이전트의 부상과 잠재력: 설문 조사": http://arxiv.org/abs/2309.07864 이를 바탕으로 두뇌, 지각, 행동이라는 세 가지 주요 구성 요소로 구성된 LLM 기반 에이전트의 일반적인 프레임워크를 제시하고, 이 프레임워크는 다양한 애플리케이션에 맞게 조정할 수 있습니다. 이어서 단일 에이전트 시나리오, 다중 에이전트 시나리오, 인간과 에이전트 간의 협력이라는 세 가지 측면에서 LLM 기반 에이전트의 광범위한 적용 사례를 살펴봅니다. 그 다음에는 에이전트 사회에 대해 자세히 살펴보고 LLM 기반 에이전트의 행동과 성격, 에이전트 사회에서 나타나는 사회 현상, 인간 사회에 제공하는 인사이트에 대해 알아봅니다.

BREAKING NEWS China Evergrande files for bankruptcy protection in New York Read more here

Productivity and Structural Reform: Why Countries Succeed & Fail, and What Should Be Done So Failing Countries Succeed by Ray Dalio In this report the drivers of productivity are shown and are used to create an economic health index. That index shows how 20 major countries are doing as measured by 19 economic health indicators, and it shows what these indicators portend for real GDP growth in each of these countries over the next 10 years. As you will see, past predictions based on this process have been highly reliable. For this reason this economic health index provides a reliable formula for success. By looking at these cause-effect relationships, in much the same way as a doctor looks at your genetics, blood tests and regimes for exercise and diet, we can see each country's health prospects and also know what changes can be made so that these countries can become economically healthier. We are making this research available in the hope that it will facilitate the very important discussions about structural reforms that are now going on and will help both the public and policy makers to look past their ideological differences to see the economy as a machine in much the same way as doctors study bodies and look at the relationship of cholesterol and heart attacks analytically rather than ideologically.