Alexander Embiricos

Everyone's talking about “founder mode,” but as it gains traction, founders will start using it as an excuse for poor management. Founder mode isn’t a license to run a company at 5,000 the same way you ran it at 5. Jensen Huang, CEO of NVIDIA, exemplifies what I like to call “scaled founder mode”. Scaled founder mode is being highly involved in the most important things, that only you can do to drive the company forward. Nvidia is one of the most successful tech companies in the world, with a $3 trillion market cap. And Huang, with a 31+ year tenure, is the longest-serving CEO in the industry. He’s a highly effective manager. And he’s still very much in founder mode. When interviewed by Stripe's Patrick Collison, Huang said “CEOs are pinch hitters. We should be working on the things that nobody else can or nobody else is.” Nvidia’s leadership team of 60 reports directly to Huang. But he doesn’t do 1:1s. This way, the team of 60 is relatively flat, with no silos and no privileged access to information. Not saying that killing 1:1s is the answer to speedy execution. But Nvidia found that this way, everyone contributes to solving problems with equal access to important info. And because Huang sees his primary role as being a custodian of the culture, he eats in the cafeteria with employees every single day. At almost 30k employees, you can’t get much more “in the details” than that.

26

3 Comments

Chat with Miles Clements about building Linear For me companies are like design problems: you have context, constraints & goals unique to you. Then you look for a solution that has a good fit to your unique problems. The company problem also evolve over times the solution Directly applying what others have done doesn’t work that well, yet there is lot emphasis on playbooks, pattern matching and playing copying what other companies are doing in industry. https://lnkd.in/ghxjJwYU

84

Insights from a deep dive into the world of Fine-Tuning LLMs! 🚀 Yesterday I had the incredible opportunity to host The GenAI Collective X OctoAI's Builders Roundtable on Fine-Tuning Large Language Models (LLMs). Here are some valuable insights we gained across three topics we discussed: 1️⃣ Fine-Tuning Foundations: Fine-tuning is particularly useful for specialized tasks such as legal or medical jargon processing, tone customization, and generating specific formats. It allows models to achieve high performance on niche datasets not covered by pre-trained models and at a much lower cost. 2️⃣ The Process of Starting with Fine-Tunes: Getting started with fine-tuning involves selecting the right data, ensuring high-quality, task-specific inputs, and carefully monitoring the model's performance. Evaluation is key to ensure your fine-tuned model meets your specific requirements. 3️⃣ Key Considerations: Effective fine-tuning requires balancing learning new information with retaining pre-trained knowledge. Watch out for overfitting and catastrophic forgetting, and always define clear evaluation criteria to measure success accurately. A huge thank you to our brilliant panelists: Aditya Jain, Benjamin Hamm, Kyle Corbitt, Sophia Yang, Ph.D., and Wing Lian. Your expertise and passion made this discussion truly enlightening! If you're interested in learning more about fine-tuning LLMs, make sure to watch the recording: https://lnkd.in/gJtSCiJu Excited for more engaging AI conversations like this in the future. 🌟 #LLM #AI #finetuning

29

3 Comments

🤖 Creating a fully autonomous AI developer to write and maintain code is still the holy grail of software development. There are many brilliant teams currently working on this, and each of them have their own unique approach. The team at Cosine have made a massive step towards making this a reality. “Cosine achieves state-of-the-art results on the SWE-bench benchmark” They have achieved a score of 43.8% on SWE Bench, which was verified by OpenAI. What's even more incredible is they have managed to do this with a lean team and a relatively limited budget. Though it’s important to note Cosine isn’t looking to replace human developers entirely, but rather augmented them with human-like assistants that they can collaborate with on any kind of coding task. Yang Li + Sam Stenner + Alistair Pullen Read more about their collaboration with OpenAI 👇 https://lnkd.in/gu7xXVPn

21

Really excited about releasing our new product "deepset Studio" today 🎉 ! We saw AI pipelines built by our users becoming more powerful, but also way more complex over the last 6 months (more components, more agentic behavior ...). No one can fit these growing architectures into their heads anymore or communicate about them quickly in a team. While we believe devs will always prefer to code a big part of their pipelines in their local IDE, the new "deepset studio" augments their building experience with a visual editor. You can seamlessly switch between both worlds as every Haystack pipeline is easily serializable as a YAML. ✅ Visually build & modify (agentic) RAG pipelines ✅ Leverage Haystack’s extensive open source library of components and integrations ✅ Import / export pipelines seamlessly to switch between your IDE and deepset Studio ✅ Deploy your pipeline anywhere: in your custom env, in deepset Cloud, or leveraging NVIDIA NIM microservices and API catalog. Join the wait list and give us some feedback on the product! We have some exciting ideas for what comes next, but will relentlessly prioritize features on what we hear from you! (links in comments below 👇)

127

7 Comments

Today I wanna share with you an amazing platform Mintlify Han Wang and Hahnbee Lee simplify developers life rapidly. How? Watch in video bellow 👇

24

1 Comment

A lot of hype around agentic AI systems, but I'm honest: If somebody starts talking about agentic AI, I stop listening. Here is why: 𝐖𝐡𝐚𝐭 𝐚𝐫𝐞 𝐚𝐠𝐞𝐧𝐭𝐢𝐜 𝐀𝐈 𝐬𝐲𝐬𝐭𝐞𝐦𝐬? Large workflows that involve LLMs to control the data flow, call tools or make plans that depend on interim results are often called agentic AI. 𝐖𝐡𝐚𝐭 𝐢𝐬 𝐭𝐡𝐞 𝐩𝐫𝐨𝐛𝐥𝐞𝐦? Agentic AI is a catch-all-phrase that basically involves any LLM-based system. People are claiming to build alls sorts of magic assistant with these stacks: Autonomous software engineers, personal travel agents and more. Because agentic AI can be everything and potentially do everything, there is a lot of hype and unrealistic expectations. 𝐇𝐨𝐰 𝐝𝐨𝐞𝐬 𝐭𝐡𝐞 𝐫𝐞𝐚𝐥𝐢𝐭𝐲 𝐥𝐨𝐨𝐤 𝐥𝐢𝐤𝐞? We are at a point where "simple" LLM-based systems such as RAG are understood well enough to be deployed in production. But even then, debugging and securing such a system is still challenging. To think that we can do this for systems that potentially involve thousands of complex LLM calls to complete a single task is ridiculous. 𝐖𝐡𝐚𝐭 𝐬𝐡𝐨𝐮𝐥𝐝 𝐰𝐞 𝐝𝐨 𝐚𝐛𝐨𝐮𝐭 𝐢𝐭? First: Let's try to talk about the technology as concretely as possible: About RAG, text2sql, code completion, tool calling etc. Please don't talk about agentic AI! Second: Let's concentrate on the (very promising) use cases that can be delivered by "simple" AI systems: Technical support bots, coding assistants, data analytics assistants, knowledge management etc. Please don't talk about magic systems that run for days to complete some hyper complex task. What do you think? Are agentic AI systems the apex of AI hype? Or is there real substance behind the paradigm? #industrialai #agenticai #hype

66

31 Comments

If you're used to monitoring backend systems, you're probably discovering that monitoring mobile apps is a totally different beast. Come learn from Virna Sekuj and Christine Hermawan how you can monitor your mobile apps better!

4

1 Comment

⚡ ion design is now live! And we’re already saving teams hundreds of hours ⚡ We turn designers into design engineers—without learning how to code. ion is the first AI app builder built for designers, transforming Figma designs into clean, production-ready code. Generate, iterate, and ship in seconds. Our goal? To create 1000x more design engineers who can: 🚀 Handle the first round of frontend development ✏️ Make design fixes, copy edits, and updates without involving engineering 🛠️ Work directly with your Git repositories ________________________ Here's how it works: 1️⃣ Design in Figma just as you always do. 2️⃣ Generate & iterate: Ion leverages your existing components, styles, and structure. Prompt it to build a working app and iterate as needed. 3️⃣ Ship faster: Create GitHub pull requests directly from ion or hand off seamlessly to engineers. We're here to empower designers, not replace them. While good design can't be automated, ion ensures you can accomplish much more and see your creations come to life—on your own terms. This means achieving more than ever without hiring additional resources. 💡 Let's make 2025 the year of the design engineer!

218

57 Comments

MOST Gen AI POCs will fail to make it to production. Why? 1. We have had some potential clients expecting us to build POCs for free ("You can put our logo on your slides, but we won't pay") - frankly if you're not willing to pay for a product, you haven't even given it enough thought or consideration. It's just wasting everyone's time. And Gen AI is not a cheap solution, especially when it's custom build. 2. There is no clear understanding of 'cost of compute'. "Let's just do a POC now, and we will figure deployment later". There's a reason why NVIDIA is a 3 Trillion $ co - there's a chip demand-supply mismatch and cost of a product (compute) in a a supply restricted market is going to stay elevated in short run. No clear deployment plan post POC is an another brilliant dead end. 3. No Data strategy - you just can't use ALL of your data on one single LLM. One of the biggest problems we have seen (and are addressing for our clients) is using PII data in Gen AI workflows. People we meet in Insurance, Banking, family office etc tell us there is more data they can't put onto Gen AI workflows, than the one they use. Perfunctory data strategy, or using simulated data to push through a POC won't get you anywhere meaningful. 4. One LLM to rule over all mistake - Way too often we come across "Lets just use the most powerful LLM and do all of this". When the fact is, some LLMs are better for some use cases you are trying to solve. Keep aside the compute cost, deployment flexibility, ownership of data etc. LLMs are tools - just like hammers and screw drivers. You can, if you wish to, hammer a nail with a screw driver, but… 5. Too many POCs - how many POCs are too many? We met someone who showed us a deck of 58 POC ideas that their company is working through (I am understating that number). I was left wondering, isn't it better to just get team leaders from different functions into a room, have a simple discussion with a whiteboard and list what problems would they like to eliminate? 58+ POCs is just lazy - the list was built by prompting chatgpt - "give me 5 dozen Gen AI POC ideas for an Bank" Gen AI is not magic. It won't make all your problems disappear over a weekend, it won't suddenly make your sales guys super heroes, or get your costs down by 50% by the time you go on your Christmas holidays. It is a super powerful tool for sure, and instead of wasting time and money on 5 dozen POCs, it might just be better to 3-5 solid ones with a clarity of how to execute till the end.

19

2 Comments

A big part of getting into YC is recommendations by alumni. Don't know any YC alumni? Work at a YC startup. https://lnkd.in/gq67hQwN If you've helped YC founders achieve an exit, you've been given the best possible exposition of the YC way. With that experience under your belt and the blessing of the founders, you'll have a strong application to YC. If you want to do a startup, I would completely skip FAANG and just work at YC startups. At a FAANG, you will condition yourself in ways that make it very difficult to do a startup, that would take tens of thousand of dollars of therapy to undo. These are divergent paths.

2

1 Comment

I brought together 5 friends from the cold email space to share their expertise. What started as casual conversations turned into learning sessions 🧠 Last time, my cofounder Enoch Chang showed you how to get started with cold email. This time, Emil Jørgensen from FinestGC share with us: - How to build a lead list - How to segment and A/B test your lead list - Different strategies for targeting companies by size - The mindset shift that drives results when selling to enterprises - How the same copy can flop in one campaign but succeed in another - How to use data to personalize your outreach - etc. Tune in here: https://lnkd.in/gzVBJMFc Cymate 🎄 p.s. didn’t expect to turn this into a little podcast, but here we are. I’d love any feedback so I can continue providing value going into 2025 🤝

24

3 Comments

📣 Today we launched a fun side project Go to http://garple.com to buy a fun domain like balses.com, exagi.com or flarse.com for your project or startup We've got over 100 domains and they are all short, easy to say, meaningless and easy to trademark. The perfect blank canvas. Just like google, twilio, quora or our own company nustom. A little bit of backstory: A long time ago, Greg Brockman came up with a technique for naming startups that involved buying the .com domain for a common english word and make that your company name - for example http://stripe.com or http://parse.com I thought I'd go the opposite way and wrote a crufty little perl script to come up with short _pseudoword_ domains, which sound like english words but aren't. I didn't have much money so I only bought ten domains, including gems such as http://orabu.com, http://vogma.com and http://exagi.com Fast forward to last summer and somebody bought http://vogma.com for $17,000! I got super excited and started writing scripts to come up with lots of similar domains names. I then got distracted, did some other stuff for a while and dusted off the scripts when it was time to name our new company, Nustom. We decided it would be fun to make some of the other domains available to other founders and makers that want to pick a decent name and just move on with things. It's a bit of a side-quest, but was great fun toasting to our first $900 in revenue with the team yesterday and listening to the Garple Garple song on repeat. https://lnkd.in/dRrmRiKp Get yours now while supplies last :D And please re-share to any startup, indie hacker and maker communities that might be interested

66

8 Comments

NVIDIA has a clear “unfair advantage” in the market right now. They are ahead of the game in so many fronts, they sit on the hardware every company working with AI needs and is the preferred partner/CVC for any AI startup at the moment. If I was Jensen Huang, I would issue 10% new stock and raise a $300bn war chest at the same time as they are doing the 1/10 split. I would take $10bn of that money, split it into 40k tickets of $250k and fund basically any new novel AI project (no ChatGPT wrappers). This way NVIDIA would have a stake in ~75% of all AI related startups and the massive amount of talent those companies attract. They would ultimately become unstoppable while still having $290bn in the war chest 🤯

23

1 Comment

As a founder, stepping back from hands-on sales and building the foundation for a team-led sales approach at Reflectiz was a major transition for the company. During the process, I learned some important lessons (and also made a few mistakes that wasted valuable time). Here’s what I discovered: 👉You know best how to sell your product As a founder, you know best how to sell your company’s product or service, especially in the early stages. And while it’s important to get ideas and directions from your sales leadership, founders shouldn't step back completely until they are confident the leadership fully understands their vision and direction. So when you first hire your sales team, stay involved. Join calls and customer meetings and don’t be afraid to challenge their sales assumptions. You know your company best. Now, I can confidently say my sales team is better than me at closing deals. But that only happened after I accompanied them through their first steps. 👉 Hire for your company’s maturity level If you’re hiring your first sales team, you probably don’t need a high-caliber CRO who knows how to scale your company from $10M ARR to $100M ARR (just yet). Instead, establish your business and sales requirements and hire someone who can take you to the next level. If you have product-market fit, a VP Sales and a CRO with a team of AEs and SDRs could be useful. If you’re not there yet, a mid-level sales manager who has experience building sales organizations will provide more value since they will be hands-on, in the trenches with you. If you’re at $1M ARR, hire the person who can take you to $10M. 👉 Get leadership boots on the ground A remote sales team needs close leadership guidance that captures your vision and energy. As a founder, either move to where your sales team is located or ensure you have a trusted leader on the ground who can transfer your passion and knowledge. If you’re bringing in a sales leader, ensuring their outlook matches yours is more important than their sales experience. Even if you’re not moving there yourself, you will need to meet the team face-to-face. Zoom alone isn’t enough for bringing remote sales teams up to speed. I made the mistake of only running weekly virtual meetings with the new team, but they were ineffective. The team lacked the full onboarding experience— learning the ecosystem, end-to-end training, hands-on knowledge transfer. It wasn’t until I physically traveled to meet them and worked with them day in and day out that they became masters of selling Reflectiz. Only then could we shift back to Zoom for ongoing guidance. Agree/ disagree? What is your experience when moving from founder-led growth to a team-led sales approach?

29

I love leaning on my fellow founder Nicolai Klemke for startup insights. It’s reassuring to know that we startup founders are in the same boat. Having someone to discuss the challenges and issues we face makes the demanding and often overwhelming process of running a fully bootstrapped startup feel less isolating. Lean on your fellow founders. #startup #startuplife #founder

6

2 Comments

Why has software remained so costly? And what happens when coding becomes practically free? Just as the internet prompted an explosion of content by reducing creation costs to near zero, LLMs have sparked a similar software boom by making code accessible to non-technical users. But while LLMs open the floodgates to a tidal wave of new software, we can’t expect all of it to be amazing. In his column in this week’s newsletter, my partner David Peterson shares his thoughts on what we might expect in this new age of software development. You can read the full story at the link in the comment section below. And be sure to subscribe to the newsletter to get exclusive insights into tech, startups, and more in your email every week.

26

5 Comments

The newest issue of The Pragmatic Engineer found a surprise uptick in eng hiring in the typically slow summer season - I'm curious to see how the macro environment impacts this uptick as we head into the fall, but I remain hopeful. I asked Gemini to give me the tldr of the article to share with my team (sorry, Gergely Orosz) and found it missed one of the most interesting pieces of the article. Gemini's highlights - Unexpected surge: Tech recruitment has seen an unusual spike in June and July, contrary to typical trends. Increased demand: Startups, especially AI-focused ones, are driving the hiring uptick. Global phenomenon: This trend isn't just limited to the US; Europe is also experiencing increased recruitment activity. Reasons for hiring: Companies are backfilling positions, investing in growth due to improved financials, and responding to increased VC funding. Challenges remain: While the job market is improving, finding qualified candidates and offering competitive compensation remains difficult for many companies. What I found most interesting - The nature of the backfills: the article finds that a lot of new headcount are actually backfills of folks that were "over-slashed" during rounds of layoffs. Not entirely surprising with the "trim the fat" and "rebuild lean" mentality, but interesting nonetheless. Either way, the uptick is encouraging (and don't always rely on Gemini's tldrs). #hiring #startups #VCbacked #softwareengineering

11

1 Comment