@WhoShivaisnot True, a lot of job postings seem fake or are just collecting resumes and for the real ones, the competition is massive, making direct applications much less effective
@MuthaNagavamsi 5 years ago, a loan felt safe because the industry felt predictable, now no one's sure what software jobs look like 1 – 2 years from now
@bidah The interesting part isn’t “Mac in the cloud”
It’s that AI agents can now operate inside a fully owned native dev environment end-to-end:
build
sign
test
ship
Mobile development workflows are about to look very different
Hi - short intro, straight to it
- Took an AI + real-time product from 0 -> 500k+ MAU
- Built low-latency systems across frontend, backend & infra
- Scaled inference (ONNX/Triton) to 1k+ RPS with batching
- Designed resilient real-time video + state recovery systems
- Operate across WebRTC, distributed systems, multi-cloud
- Strong systems mindset: performance, observability, safety
- Founding eng - shipped product + infra end-to-end
Looking to build and scale something ambitious, happy to connect for an opportunity
– Ayush
Portfolio : https://t.co/7uv15XkW2E
@k1rallik one of the ways to increase ARPU (Average Revenue Per User), disguised as a tokenizer update + silent defaults + “bugs” that all conveniently increase usage
@iddybrown1 Hi - quick note, not a pitch
- Building AI + real-time systems at scale (0 → 500k+ MAU).
- Spend most of my time on low-latency infra + inference.
Feels like there could be overlap, worth a quick chat?
@signulll Is it actually bad or just not optimized for real coding workflows?
Feels like a big difference people aren’t calling out. Curious what’s actually breaking for folks
Still shipping with it, our backend at Vooz runs on C#
From what I see, it’s fast, reliable, and way more modern than people give it credit for. The “old tech” narrative feels pretty outdated at this point.
Genuinely curious, what are people replacing it with that’s actually better, not just trendier?
@TheAhmadOsman This is a solid setup 👀
curious though, how’s the context window holding up in practice when you’re running this locally?
Are you hitting limits with longer chains / agent workflows, or does it stay pretty usable with that stack?
Would love to hear real-world experience