Mathias B
324 posts

Mathias B
@mathiasvrb
Grew game servers to $25k MRR. Switched to saas after watching them drop to $500 MRR
เข้าร่วม Haziran 2025
266 กำลังติดตาม89 ผู้ติดตาม

day 35 building saaS in public.
waitlist just crossed 1000 people.
started texting a few of them directly on whatsapp to understand what they actually need before we launch. way more useful than guessing.
turns out most of them just want one thing: stop spending hours making tiktok content manually.
back to building.


English

Hey @X
I want to connect with people interested in:
Cybersecurity
Ethical Hacking
Tech & Backend Dev
Java
TryHackMe / CTFs
Linux
Building in public
Drop "Hi" and let's connect 👇
English

New Post: Digital Marketing and SEO news updates 2026-06-30 thewealthywalk.com/digital-market…
English

I just published The Best Vibe Coding Tools in 2026: Move Beyond Basic Chat Prompts medium.com/p/the-best-vib…
English

20 more for 300!
Looking to #connect with more people in tech/build in public 👋
Especially from:
- Web/App development
- AI/ML
- Indie Hackers
- Startups
- Growth & Marketing
- Data Science
- Product Design / UI UX
- Tech in general
Let's connect and grow together ⚡

English

@oldstackjournal building an AI growth co-pilot for X, happy to connect :)
English

@CyberSec_Queen And I bet they sometimes offer superior customer service lol
English

Every team shipping a coding agent — Claude Code, Codex, Cursor — is really running a serving-systems problem. The "tech behind the tech" is the LLM-serving stack underneath, and until now nobody had real data on what that workload looks like.
New arXiv (2606.30560) from @bariskasikci's SyFI lab (@UWSyFi, @uwcse) is the first large cross-provider trace of real coding-agent use: ~4,300 sessions, 350K LLM steps, 430K tool calls, 43 developers, 8 months, Claude Code + Codex.
It breaks the intuition that agents mean long generations. The median step replays ~119K context tokens to emit just ~214 output tokens — two orders of magnitude more reading than writing. So the bill is the context, not the answer: prefix tokens are 59.5% of total cost.
Tool calls are brutally long-tailed: 80+ tools, but the top 3 are 80%+ of calls, and the 4% of calls that run >1 min eat 85% of all tool time.
And the prefix cache everyone leans on? 95.7% hit rate — yet misses cluster right after a human pauses to think, amplifying prefill 3.8x. Those human-gap misses alone are ~46% of fresh tokens and ~13% of spend.
For technical leaders: your agent's cost and latency live in the loop, the replayed context, and the idle gaps — not raw token generation. Tune tool-call overhead, append-length-aware prefill, and KV-cache eviction around human gaps before you scale the fleet.

English

@Bayrus_Consult Especially with how claudes been acting the past 2 weeks 😠
English

The fastest way to lose an AI user?
Make them wonder if today's result will match yesterday's.
#SaaS #UX #Activation
English

@netcentrics_UK True true, I recommend the book "The SaaS Playbook" by Rob Walling
English

"Bootstrapping" means self-funding your startup, from "pull yourself up by your bootstraps" - which originally ridiculed the impossible. Self-funding isn't impossible, but don't expect it to be easy or cheap. Success tends to be proportional to investment. #SmallBusinessTips
English

@MB_BlueLogic 80% sounds about right, support especially is basically solved now
English

Most solopreneurs hire support first, then marketing, then ops. AI tools now handle 80% of that for less than you spend on coffee.
logicimpactai.com/blog/2026-06-2…
#Solopreneur #AI
English













