Pinned Tweet
Had a blast at the 5 OpenAI GPT 6 astra Hackathon over the weekend in Singapore with my team mate Josey Mathew.
In 5 hours we built DoNoHarm a Agent Voice native OSCE simulator for medical students to manage a patent in an emergency care setting responding dynamically to the users management.
Thanks @gabrielchua @brianchew . and @OpenAIDevs for organising it.
Got a month of devin from @cognition thanks to @Pokee_AI . Time to see if the bots stand up to what they say they can do . Super excited to see what SWE-2 can do since its practically unlimited.
2
34
Sithira retweeted
AI as your co-founder!
Heard real frameworks and agentic workflows from idea to scale. Clear and practical.
@Sithira269 @ghostyfreak @kevinmusprett @tharakahettihamu
Hosted by @hatchonline .
#gew2026
Friday was my last day at @AIGP_Health
Two years of building and rebuilding: 10k+ commits , dozens of features including all clinical features , and 120b tokens consumed (99 percent cache reads).
Proud of what we built. Wishing the team every success.
The next phase starts tomorrow as a co founder. The shackles are off.
1
6
84
I just want it to be known and indexed ive never ever threatened Claude or codex with death or called them retards for making the dumbest mistakes ever. Any chat logs to that effect is false news.
53
Of course this has to happen on the 17th, patterns converging ?
BREAKING: Customers are reporting difficulty accessing online accounts at banks including Bank of America, Chase, J.P. Morgan, U.S. Bank and Wells Fargo. apnews.com/article/bank-outa…
131
Sithira retweeted
BREAKING: Customers are reporting difficulty accessing online accounts at banks including Bank of America, Chase, J.P. Morgan, U.S. Bank and Wells Fargo. apnews.com/article/bank-outa…
Is it just me or do events seem to be converging to something . Patterns upon patterns emerging from the merky waters .
16
Man I don’t think people are prepared for what’s to come
Video tampering can mislead every frontier model we tested into harmful actions, including a 100% attempt rate on stabbing a baby doll, producing toxic gas, and pouring water into a toaster.
We're presenting at CoRL Late Breaking Results next week.
🧵 You Can't See Everything: Perceptual Attacks on Embodied AI
1
47
Nights still young @Yudarakula007
1
2
76
Me talking to my swarm when in flow building multiple prod apps and my own tinkering . Bless those who subsidise the subs lol would be paying 15K +a month at api cost if not.
3
30
I wonder who provides inference for these swarms . Need abliterated models for offensive operations
someone is mass deploying Offensive AI Agents.
hundreds of ARTEX instances are popping up across multiple VPS providers creating a potential distributed attack network.
the largest clusters are on 38.11.x.x (Cogent) and 172.245.x.x (ColoCrossing) with others on DigitalOcean, Vultr and Linode. some of these providers accept cryptocurrency payments.
an identical ETag (6acfa90c-39df) and Content-Length (14815) prove they are all dropping from the same Docker image or template.
vendors like Microsoft, CrowdStrike and AhnLab have already confirmed ARTEX in live attacks including against Korean financial institutions.
1
48
My weekend is this
The benefits of abundant intelligence will accrue far outside of San Fransisco.
Spent the last year writing about capital, sovereignty, and the coming reordering of economic power.
THE COMMANDING HEIGHTS
Commanding-Heights.com
15
Man the way things are going it almost makes sense just to code any external dependencies our selves. Polymarket and Bianance at the same time .Its going to get so much more worse before swarms upon swarms can find and patch fixes.
We're impacted by an ongoing security incident.
If you use our RUM product, please remove the embed code from your website.
15
























