Max Romanovsky

958 posts

Max Romanovsky banner
Max Romanovsky

Max Romanovsky

@M_Romanovsky

Head of AI and SWE | Writing about AI, startups and vibe coding | Opinions my own

Berlin, Deutschland Katılım Eylül 2023
629 Takip Edilen194 Takipçiler
Sabitlenmiş Tweet
Max Romanovsky
Max Romanovsky@M_Romanovsky·
🧵In December, I and my friends ran a survey across developers about vibe coding and AI tools. 👥 140+ participants 🧑‍💻 Most common roles across participants: Team/Tech Lead (24%), Full-stack (19%), Backend (14%), Founder (11%) Key insights in this thread:
English
5
2
8
631
Christos Zoukos
Christos Zoukos@cz4000·
@M_Romanovsky Still have a way to go but we getting there! We still need a cultural change on entrepreneurship and risk taking though
English
1
0
1
8
Max Romanovsky retweetledi
Max Romanovsky
Max Romanovsky@M_Romanovsky·
The math behind why most founders fail Let’s model a startup founder as an idea-testing system. Assume the probability of any single business hypothesis being a power-law winner is p = 0.02 (2%). This is generous - most VC data suggests it’s closer to 1%. If you test n = 1 hypothesis (which is what most founders do), your probability of success is: P(success) = 1 - (1 - p)^n = 1 - 0.98^1 = 2% Now scale n: ∙n = 5 → P = 1 - 0.98^5 = 9.6% ∙n = 10 → P = 1 - 0.98^10 = 18.3% ∙n = 25 → P = 1 - 0.98^25 = 39.7% ∙n = 50 → P = 1 - 0.98^50 = 63.6% ∙n = 100 → P = 1 - 0.98^100 = 86.7% Read that again. Going from 1 to 50 hypotheses takes you from a 2% chance to 64%. Same founder. Same skills. Different strategy. Now add the expected value layer. In a power-law distribution, the winning outcome isn’t 2x - it’s 100x-1000x. So the EV equation looks like this: EV = P(success) × V(power-law outcome) - C(experiments) If V = $10M and each experiment costs $2K and 2 weeks: ∙n = 1: EV = 0.02 × $10M - $2K = $198K ∙n = 50: EV = 0.636 × $10M - $100K = $6.26M The cost of experimentation is linear. The expected payoff is exponential. This is the entire argument. Why most people get this wrong They optimize for depth instead of breadth. They spend 12 months working on one idea instead of spending 1 month testing 10. The sunk cost fallacy compounds - every extra month invested makes it psychologically harder to kill a bad bet. The power law is not a metaphor. It’s a probability distribution. And it has a direct mathematical consequence: the optimal strategy is to maximize n (experiments tested) while minimizing C per experiment (cost per test). In practice this means: ∙Set a kill threshold before you start (e.g., “if no signal in 2 weeks, next”) ∙Define “signal” quantitatively (X signups, Y revenue, Z engagement) ∙Budget time and money per experiment, not per “startup” Test more. Kill hypothesis faster. Grow the winning idea.
English
1
1
6
53
Max Romanovsky
Max Romanovsky@M_Romanovsky·
How many startup hypothesis did you test in past 12 months
English
2
0
4
52
Viktor Seraleev
Viktor Seraleev@seraleev·
Indie developer superpower: Speed. No meetings. No approvals. Just ship.
English
6
3
50
2.1K
Max Romanovsky retweetledi
Sam Altman
Sam Altman@sama·
GPT-5.4 is launching, available now in the API and Codex and rolling out over the course of the day in ChatGPT. It's much better at knowledge work and web search, and it has native computer use capabilities. You can steer it mid-response, and it supports 1m tokens of context.
Sam Altman tweet media
English
2K
1.2K
12.9K
1.3M
Max Romanovsky
Max Romanovsky@M_Romanovsky·
The most important advice I can give to everyone who uses AI at work: never send AI-generated output to your manager or to your team without reviewing it first. Nobody wants to read another piece of AI slop in your performance review form, documentation or an incident analysis. People are not paid to copy-paste from ChatGPT, and nobody cares that it was not your mistake - the AI generated it that way. If you send something under your name, it is your responsibility and your reputation. This does not mean you should not use AI - just spend the time to verify it, fix mistakes, and improve the result. Want me to adjust the tone (more casual, more formal, shorter), tailor it to a specific industry, or turn it into a different format like an article or email?
English
0
0
2
43
Max Romanovsky
Max Romanovsky@M_Romanovsky·
One of the first layoffs where AI is clearly articulated as a reason
jack@jack

we're making @blocks smaller today. here's my note to the company. #### today we're making one of the hardest decisions in the history of our company: we're reducing our organization by nearly half, from over 10,000 people to just under 6,000. that means over 4,000 of you are being asked to leave or entering into consultation. i'll be straight about what's happening, why, and what it means for everyone. first off, if you're one of the people affected, you'll receive your salary for 20 weeks + 1 week per year of tenure, equity vested through the end of may, 6 months of health care, your corporate devices, and $5,000 to put toward whatever you need to help you in this transition (if you’re outside the U.S. you’ll receive similar support but exact details are going to vary based on local requirements). i want you to know that before anything else. everyone will be notified today, whether you're being asked to leave, entering consultation, or asked to stay. we're not making this decision because we're in trouble. our business is strong. gross profit continues to grow, we continue to serve more and more customers, and profitability is improving. but something has changed. we're already seeing that the intelligence tools we’re creating and using, paired with smaller and flatter teams, are enabling a new way of working which fundamentally changes what it means to build and run a company. and that's accelerating rapidly. i had two options: cut gradually over months or years as this shift plays out, or be honest about where we are and act on it now. i chose the latter. repeated rounds of cuts are destructive to morale, to focus, and to the trust that customers and shareholders place in our ability to lead. i'd rather take a hard, clear action now and build from a position we believe in than manage a slow reduction of people toward the same outcome. a smaller company also gives us the space to grow our business the right way, on our own terms, instead of constantly reacting to market pressures. a decision at this scale carries risk. but so does standing still. we've done a full review to determine the roles and people we require to reliably grow the business from here, and we've pressure-tested those decisions from multiple angles. i accept that we may have gotten some of them wrong, and we've built in flexibility to account for that, and do the right thing for our customers. we're not going to just disappear people from slack and email and pretend they were never here. communication channels will stay open through thursday evening (pacific) so everyone can say goodbye properly, and share whatever you wish. i'll also be hosting a live video session to thank everyone at 3:35pm pacific. i know doing it this way might feel awkward. i'd rather it feel awkward and human than efficient and cold. to those of you leaving…i’m grateful for you, and i’m sorry to put you through this. you built what this company is today. that's a fact that i'll honor forever. this decision is not a reflection of what you contributed. you will be a great contributor to any organization going forward. to those staying…i made this decision, and i'll own it. what i'm asking of you is to build with me. we're going to build this company with intelligence at the core of everything we do. how we work, how we create, how we serve our customers. our customers will feel this shift too, and we're going to help them navigate it: towards a future where they can build their own features directly, composed of our capabilities and served through our interfaces. that's what i'm focused on now. expect a note from me tomorrow. jack

English
0
0
0
61
Max Romanovsky
Max Romanovsky@M_Romanovsky·
Speaking at CDO BFSI exchange today
Max Romanovsky tweet media
English
0
0
2
49
Alex Xu
Alex Xu@alexxubyte·
10 Types of API Testing
Alex Xu tweet media
English
11
94
499
21.6K
Max Romanovsky
Max Romanovsky@M_Romanovsky·
@cyntro_py The same with warranty, I used ChatGPT several time for cases when I need to return or repair something. 100% success
English
0
0
0
199
stepan
stepan@cyntro_py·
I had money stuck with a crypto card provider. Months of support tickets led to nothing. Claude Code sent one email threatening regulatory action which resolved it in 2 days without any actions from my side. Today, there's zero marginal cost to fight back. Every grievance gets filed by customers.
English
16
4
51
6.2K
Tibo
Tibo@thsottiaux·
We’ve made GPT-5.3-Codex-Spark about 30% faster. It is now serving at over 1200 tokens per second. More to come on speed across the board.
English
216
120
2.6K
348K
Max Romanovsky
Max Romanovsky@M_Romanovsky·
@tegmark We are in extended version of prison dilemma. One of our problems that all players have different estimation of their reward function. Looks like there is not easy solution for that and the equilibrium is in negative values.
English
0
0
1
59
Max Tegmark
Max Tegmark@tegmark·
Thanks Dean for saying the quiet part out loud! I'd add that if US companies are allowed to continue racing toward AGI/ASI with zero restrictions, the US will also functionally cease to be a republic: our government will de facto be overthrown first by an AGI company and then perhaps by its machines.
Dean W. Ball@deanwball

I don’t want to comment on the DoW-Anthropic issue because I don’t know enough specifics, but stepping back a bit: If near-medium future AI systems can be used by the executive branch to arbitrary ends with zero restrictions, the U.S. will functionally cease to be a republic.

English
44
44
312
42.2K
Bela Wiertz
Bela Wiertz@blwiertz·
>600 applications for only 80 seats... That's how the reality of @techeurope_ hackathons looks like! 1 year ago, we started with a simple mission: We want more big technology companies to be built in Europe. There are so many things to get right to make that happen, and hackathons are only a small part and a lot of very draining groundwork for an ecosystem. But what I've seen in the last year made me believe more than ever! Congrats to the winners of our Paris Hackathon at Neon Noir: Bastien Mirlicourtois, Anthony Akinboh, Wei-Ling HUNG & Luli Maruyama
Bela Wiertz tweet mediaBela Wiertz tweet mediaBela Wiertz tweet media
English
6
4
41
2.4K