Joseph Emison

PSA: If you send a founder product nits + feedback with concomitant screenshots/mocks you instantly vault into being an S-tier user. (Friction logs are even better, of course.)

2

#Fintech organizations have 3 main NEVERS: NEVER want to see downtime NEVER default on regulations NEVER miss a chance to know their customers But achieving this at scale = Ingesting TBs of data = spending millions of $$$ on SaaS observability tools = sharing financial/PII data with vendors = compliance issues This simple yet straightforward (+ extremely costly) gap This is where Incerto’s Observability solution comes in = Single unified platform, custom-built to your infrastructure = 85+% guaranteed cost savings + 100% control = #OwnYourObservability Still on the fence on making the switch from SaaS observability? You’d surely want to hear from our clients from FYERS, Dhan, and other leading orgs. #observability #financialservices #monitoring

22

Generative AI is most useful to corporations when they define use cases and have Subject Matter Experts in those areas provide their own data to the tool (in a secure environment) and iteratively prompt to create the best assistant they ever had. Imagine if an underwriter writing complex P & C contracts for properties tied to many other properties. He/she has to consider many contracts, localized weather data, and many other sources of information. Tell the LLM: “Look at the attached examples of subrogation clauses we endorse for our clients and tell us how the clauses in the contracts associated with this property differ from those.”

3

3 Comments

"Falling costs of building new software companies are a mirage, and the reason for that is this thing in economics called the Jevon's paradox... ...the Jevon's paradox occurs when technological progress increases the efficiency with which a resource is used (reducing the amount necessary for any one use), but the falling cost induces increases in demand enough that resource use overall is increased, rather than reduced." Marc Andreessen with a counterintuitive take on what the future could hold for the software industry. #tech #technology #innovation #software #deeptech #startups #entrepreneurs #entrepreneurship #ai #artificialintelligence #vc #venturecapital #money #business

510

60 Comments

For those who remember my previous post about the gov(dot)br hack, check out the HackNotice data for the dropbox domain. Leaked records on the domain are way above average. 25 is really bad and you should expect an incident like the current one if it gets that high. If you're responsible in any way for an organization, you've got to be keeping an eye on this stuff. When ZeroHack tells you a hack is coming, it's because it is. Normalize seeing around corners and preventing incidents, instead of getting caught flat footed doing damage control.

3

It feels like every few weeks there's a new model release exceeding some predetermined set of benchmarks every time (MMLU, Math, GSM8K, etc.). Quite frankly, it feels like a lot of these benchmarks aren't great proxies for what users and companies perceive as model performance. BUT, the LastMile AI team has been testing Claude 3.5 Sonnet for coding and it definitely feels different; noticeably better for coding than GPT-4o. Have others tried Claude 3.5 for coding yet? Curious to hear others' experiences.

41

12 Comments

The journey from $1 to $10M in ARR is typically an exercise in finding your first channel-market fit as a GTM/Marketing leader and/or founder. After prioritizing the top 1-3 channels you believe are most likely to be a fit for your product, finding channel-market fit is typically an exercise in experiments and iterations. If you’re flying blind bc you don’t have the systems, processes, and data hygiene setup to measure and test quickly inside a framework, it’s going to be a difficult journey. Check out the podcast I did with the team at Evron for more mistakes to avoid on your journey from $1M to $10M ARR.

41

5 Comments

In 10 years, I bootstrapped $0-1M ARR twice, and just 14 weeks in we’re at $800k ARR with RB2B. Here's the highly controversial secret weapon I used to go $0-1 (and why it might NOT be right for you): SHITPOSTING Here’s a breakdown of what I did, why I did it, and whether you should do it to: Shitposting Campaign #1: The GetEmails (now Retention.com) Facebook Ads Campaign of 2020 - I made 1min talking videos answering the sales questions w/ Helen (now wife) - They were amusing, we used props, we rotated backgrounds - A month in I realized that I was agitating a vocal minority of privacy people - They were calling me a scumbag for reasons X, Y, and Z  - I made ads specifically about X, Y, and Z, and their heads exploded - Facebook loves this more than anything, the ads crushed - People still say they recognize me from those ads today!!! Results: $0-1m ARR in 26 weeks. 6 FTE. Shitposting Campaign #2: The 6Sense Saga of 2024 - I was upset at a slimy tactic 6Sense used to get us to renew early - I put up an inflammatory (and hilarious!) post about how they are afraid of me - By some act of God, they sent me a Cease and Desist - I posted the Cease and Desist and re-wrote the shit post  - It was the single greatest PR campaign for any 8 week old startup ever - Everybody on LinkedIn hates 6Sense, so it lit a fire I NEVER expected Results: $0-1m ARR in ??? weeks. 5 FTE. (I can’t tell right now but if I had to guess I’d say it will be 17 weeks). TAKEAWAY Should you shitpost? On one hand, you might look at this and say, “social media loves it, so DEFINITELY”. You would be correct about social media loving it. The problem is you run the risk of ruining your reputation… So if you do, it BETTER be worth it. Separately, I’ve found that it’s also very bad for my mental health. When I’m shitposting, I’m living in a world of negativity that I do not thrive in. My wife says I turn into the worst version of myself. I’d say don’t shitpost unless you are: - At a VERY early stage - There is an INCREDIBLE awareness opportunity I’m going to be going into great detail about my own shitposting campaigns at the Inbound-Led Outbound workshop Tuesday at 3pm ET. We’ll break down: 1. How to shitpost  2. The numbers of shitposting 3. When you should/shouldn’t Register here: https://lnkd.in/gs-Yyc4p

332

119 Comments

Came across this on Hacker News last night, definitely worth the read for anyone in our industry. We’ve all been watching the train wreck that Google Search has become in recent years and this article does a good job of outlining what happened on the inside. I’ll be touching on some of this in my talk tomorrow but this whole escapade is one of the reasons that getting accurate data on the SERPs at as broad a level as possible is so critical to any site that depends on search traffic, paid or organic. The use cases are numerous, including cannibalizarion detection, competitive opportunities, ad monitoring, copy trends, SERP layout shift, local changes, and much more. https://lnkd.in/gtE5j2Ez

8

6 Comments

SaaS is DEAD...... Of course it is NOT. #AI will replace it all. Not quite as a generalized statement. What a fantastic episode that dropped last week from Harry Stebbings with one of the best AI leaders out there Alexandr Wang So much to cover there but the primary thing from me was data for models has hit a wall and now "Frontier" data as Alex says is what is needed. Deep, complex, chain of thought reasoning tasks and so much more. Want to slap a chatbot UI on your app? Tablestakes at this point. It isn't to make it work properly and accurately but the simple function is. Ensuring you are becoming agentic is the new oil. Go read Raghvender Arni last few posts on this. A quick visual below from crewAI as they are building something pretty cool. Agents are going to make so much more possible. Be agentic so you win. So what does all of this mean for GTM? A lot....here are 5 key items to start: 1. You can now finally build an architecture that can sell "outcomes". If you read what Aaron Levie wrote this past weekend on this relative to seat based licensing you will see a massive TAM explosion. Everyone wants to "sell" a hard ROI based on an "outcome". Everyone also wants to "buy" that. But has this really been true? Not in a widespread manner. This is changing fast - so better the pricing model (seats or otherwise). 2. Content - Can't just explain why your AI is better. Why you are delivering the outcomes? I met with someone doing some wild things in Oil and Gas last week. Deep domain expertise delivering 10s of millions of hard ROI and 7 figure deals to match. Infuse your domain, process, architectural and marketing experts together and DO NOT use "just" AI to produce it. Remember - if you sell outcomes, you get more money. If you get more money on outcomes, the customer gets more value. This is a winning relationship. Both sides deserve this. We are here. 3. Systems to manage GTM in this world. Forget the myraid of ways having an agentic GTM "stack" would be amazing (finally we don't have to rely on an expensive database called the CRM no one likes) but how do you use all of this information to get the right "tiggers" for your ICP and personas? Much more available to us now. If you don't use this automation you will lose. 4. GTM Strategy. Perhaps this will involve some value added services. Some hate this. Some love it as it makes you sticky. If it is high margin, high value and delivers the outcomes - does it matter? Is it just pre-sales and part of your S/M expenses? Maybe but then the personas you hire are different. Lots to ponder - no right or wrong in any of this. 5. Recruiting - In the next evolution that what is your talent strategy? Some nail this now. Some use amazing partners. Investors are amazing at intros but you won't be able to rely just on that as competition is fierce and many of these "roles" are changing as we speak. Same old playbook won't work. Happy Monday.

9

1 Comment

Bootstrapping is better for founders and employees 🤝 Aligned Incentives: Founders and employees share a common goal for sustainable growth and success 🏆 Long-Term Focus: No pressure for rapid exits and hiring for the sake of a large head count 📈 Ownership: Fewer dilution rounds = larger share of the company’s equity What do you think?

46

4 Comments

“If we don’t start, we’re going to be left behind.” I had the chance to chat with Chris Prodoehl of Chicago Tube & Iron and he shares the real thought that’s on a lot of business owners minds when it comes to AI. That is, where do I start? For many, implementing AI into your business might feel like a daunting task. But the reality is, it doesn’t need to be. Focus on the processes that are valuable to your business today and identify the obstacles that could be more efficient with the use of artificial intelligence. And, ensure you have a trusted partner guiding you along the way. Artificial intelligence is here to stay, and sometimes the best way to get started is by experimenting. Epicor #ERP #AI #ArtificialIntelligence #CustomerSuccess 

96

1 Comment

Google is failing at supporting companies through their email changes. Postmasters sucks. Plain and simple. It gives you extremely limited data, its methodology is opaque, and its domain & IP reputations are misaligned with actual email deliverability more often than not. The worst part? They don’t feed that data back to the email service providers like other inbox providers, meaning there is no deterministic way to understand your spam complaint rates, or why you may be landing in spam. Given Gmail makes up ~65% of a typical B2C email list, they have a responsibility to help brands navigate these changes. The big companies will be fine. They have expert email marketers and teams of data analysis who will test into the right segments & send cadence. Who this really hurts is the SMBs that rely on email for 30% of their revenue. They need to do better.

1

3 Comments

Love it or hate it: It’s all over the news. Today is GameStop’s Annual Shareholder Meeting. 🕹️ And what that means is that Urvin’s Verified $GME Shareholder Community is going to be BUZZING! 🐝 Urvin’s Verified Shareholder Communites are a safe, secure way for shareholder communities to engage with one-another knowing that they are surrounded by real, confirmed, shareholders - without concern for Bots, Shills, and other counterproductive elements. 🤖 How? Urvin’s allows users to connect their brokerage portfolios - for free - to (1) personalize the data and news feeds for the entire site, yes - but really (2) gain access to exclusive verified communities filled with other connected and verified user accounts holding the same equity (or other asset) as you. 🔐 Why does it matter? Because social media is broken. Bots account for 60% of all traffic by some estimates. We’re all being force-fed fake and inauthentic messaging. What you’re seeing isn’t real, and with AI it’s going to become increasingly unreal - and not in a good way. To clear the air and reconnect, a new beginning is needed. A new social environment filled with authentic, verified voices. Somewhere where people, shareholders can speak to companies, and companies can speak to shareholders - online - without fear or distraction. That’s what Urvin is. Take a look. 🟣 #shareholder #shareholderrelations #startup #startups #investors #investor #corporaterelations #founder

28

1 Comment

🤯 On average, young companies (1-2 years old) start out with 29 SaaS apps. By the time they’re 3-6 years old, that number spikes to 103 apps. 📈 We're using more apps than we ever have. This is contributing to major context switching. Some studies estimate that people spend as much as 35% of their time looking for information. When we started building the newest iteration of Hoop we wanted to help people with the problem of: "I feel like I'm constantly missing something. Information is spread across too many tools. I spend more time managing/looking for tasks than actually working." Hoop automatically captures tasks across Slack, Google Meet, and Zoom. It then ports them into an easy to manage, familiar to-do list user interface. If this sounds intriguing, join the waitlist today. I'll be happy to walk you through the product personally. 😊 https://lnkd.in/e_Nmbr93

2

You can craft better agentic prompts if you ask yourself: What would you tell a task contractor if they can't email you back? Write LLM prompts with detailed task context, relevant information, instructions, and most importantly your preferences

2

1 Comment

My teams have shipped >100 LLM use cases over the past 12 months for F500 companies and startups. Here are four simple metrics you can use to evaluate whether your RAG is healthy: 1. Compare the question and the source.  Ask: Given the question, did we pull in a relevant source or context?  This is called Context Relevance. 2. Compare the question and the answer.  Ask: Is the generated answer relevant to the question being asked?  This is called Answer Relevance. 3. Compare the generated answer with the correct answer.  Ask: Do they match? This is called Answer Correctness. 4. Compare the generated answer with the source.  Ask: Can the generated answer be backed up?  This is called Faithfulness or Hallucination. What other RAG topics are you interested in? For RAG experts out there, what are your greatest evaluation wins and challenges? #RAG, #AI, #MachineLearning, #EvaluationMetrics, #ContextRelevance, #AnswerRelevance, #AnswerCorrectness, #Faithfulness, #Hallucination, #AIModelEvaluation, #AIModelHealth

12

As I educate myself in the venture capital realm, I learn about the ups and downs of the startup world. The Lacework news is interesting because when a startup is overvalued, it means its market price is higher than its true business value. If it's later bought for much less than its inflated valuation, it indicates that the market has corrected its perception of the company's worth. This situation can be tough for the original investors and founders who may face significant financial losses, as they won't get back as much money as they expected. It can also affect the startup's reputation and the morale of its team, as such a stark adjustment in value can lead to doubts about the business’s viability and future prospects. Not sure if this is true for this situation, but I have been told it occurs more often than not. #VC #stealth #startups #cloud #cybersecurity #VentureCapital #learnNewThings

5

1 Comment