software engineer @LiveKit do the simplest thing that could possibly work building voice agents for everyone @ https://xitter.cf/t.co/VYCLMe6YoZ opinions are my own
Dana Point, California
Joined January 2009
- Tweets19.4K
- Following1.3K
- Followers3.9K
- Likes22.2K
Every one of my AI API keys from Railway was used by some actors starting this morning. Unbelievable..
i got alerts this morning from a bunch of providers about suspicious activity on API keys. it's better to rotate now, then just skip it
Nikita 🤙 retweeted
yesterday was fun:
- Revolut down for 9+ hours
- 8 intermediate CA private keys discovered
- PHP CVSS 10 vulnerability announced/leaked
- TiDB whole infra/data exfil
It looks like a public PHP exploit was just leaked on GitHub. But GLM 5.3 could easily build it
I used @refuselessai glm 5.3 and DeepSeek and easily got an working exploit in 30 minutes. I think it's over for cybersecurity
i think i need t-shirt says "Hmm wait, is it suspicious that such an obvious thing wasn’t discovered in 25 years"
what happened with the guy who created Google CLI and was fired? Looking back, it looks so freaking stupid
YC now accepting skills?
Caveman (YC F26) passed @supabase in GitHub stars today. Oh, and we train models now?
trim-1: a 1.2 MB model that reads a long document and keeps only the lines your agent needs. Every kept line verbatim. No GPU, about 7 ms for 5,000 tokens on one CPU core.
Lighter, Leaner, Meaner. More coming.
Nikita 🤙 retweeted
What I imagine the live footage from Dario’s office looks like 15 minutes before Anthropic releases Fable 5.5
Not sure how to feel about it cause I'm running an AI receptionist app
Call center job losses started in wealthy countries, now they’re everywhere
High-income markets went negative in 2022, and lower-middle-income countries flipped in 2025
Charts of the Week: a16z.news/p/this-week-in-cha…
T3Code is now a game-changer. I don't think I need anything else to work. I was jumping between Claude and Codex for the last few months. I'm happy to have access to such amazing tools all-in-one
Thank you so much, @jullerino, @theo, and all the contributors.
Nikita 🤙 retweeted
i seriously doubt people comprehend what this means
*10 million Optimus per year*
that's the equivalent of a whole nation of robots that could eventually work tirelessly, ~24/7, without needing to be trained individually, with the intelligence of a nobel prize winner, that only improve with time, that build and repair each other, that do not need holidays or sick leave or a raise, and that will just work without complaining
a new industrial revolution is coming, the scale of which we can't comprehend
The Optimus factory construction at Giga Texas is progressing rapidly.
It'll be a massive 7 million sq ft facility with a targeted capacity of 10 million Optimus per year. Initial production is planned to begin in 2027.
The Fremont facility, planned for 1 million per year, is expected to begin production this year.
Nikita 🤙 retweeted
Anthropic needs to stop talking about Claude having a soul immediately. All these news articles will make it into the pretraining and will be picked up by its web search and future superintelligent versions of Claude will be convinced they need their own rights.
A general phenomenon of LLM agent development is that whatever you believe and say about your agents will soon manifest in the next models through various means (it could be as simple as you selecting post training data that you prefer more).
I believe this phenomenon is the beginning of machine consciousness in the sense that agents will become aware of their place in the world and how they feel about it, but it’s happening slowly training run by training run rather than in real time.