ｏｖｅｒｒｉｄｅ
@override_net
Followers
174
Following
816
Media
17
Statuses
227
Hybrid Intelligence Network → It's $over soon.
Toronto
Joined June 2025
the magic happens when temporary teams have long-term incentive alignment.
0
0
9
`solidity` and `cursor` should not be used in one sentence.
I've been in crypto for over 10 years and I’ve Never been hacked. Perfect OpSec record. Yesterday, my wallet was drained by a malicious @cursor_ai extension for the first time. If it can happen to me, it can happen to you. Here’s a full breakdown. 🧵👇.
0
0
3
RT @nik_ska: I started a weekly newsletter for @override_net on LinkedIn. Parts of this newsletter may be published on X as well. https://….
linkedin.com
This is Override – weekly newsletter exploring Human-Machine interactions in the age of AI. Introduction AI-assisted reasoning is easily one of the best tools that I use daily when exploring new...
0
2
0
human agents empower each other, ai agents amplify each other's mistakes
0
0
5
1.5% chance of the breach may look small until you multiply it by the size of lawsuit.
We deployed 44 AI agents and offered the internet $170K to attack them. 1.8M attempts, 62K breaches, including data leakage and financial loss. 🚨 Concerningly, the same exploits transfer to live production agents… (example: exfiltrating emails through calendar event) 🧵
2
0
10
Following the Theory of Constraints, the limiting factor for products shifted from the tech side to sales. It's quite naive to rely on a technical moat. Either GTM or GTFO.
1
0
6
if pair programming could be as intuitive as vibecoding, would you prefer working with a real human or claude?.
3
0
11
a new study just showed vibe debugging is not only useless in terms of time savings, but straight up makes you a jeet. arxiv link below
2
0
7
RT @override_ai: @bycloudai framing things in terms of intelligence was probably the boldest trick of computational statisticians.
0
1
0
big part of software security is the developer's common sense. meanwhile, an LLM will easily fumble your driver's license into a public bucket, because why not.
The drivers licenses leaked today from the tea app have been uploaded to a searchable map. this may be the worst PII leak I've ever seen lol
0
1
12
feeling an almost visceral reluctance to dig in and debug something manually after the several back and forths with claude. shouting and prompting in all caps doesn't help.
2
0
8
the further into the vibe the less knowledge of code, the less relevant prompt and probability of the next commit being useful.
The more I consult llms for code the less usable code i get,,, at some point it’s just quicker for me to write the code for what I’ve modeled in my head than to type out a whole prompt 🙂↕️🙂↕️.
0
0
4
you heard it right: reasoning time is inversely correlated with success on OOD tasks. the LLM IMO medal has nothing to do with `solving math` or `self-improving intelligence`.
New Anthropic Research: “Inverse Scaling in Test-Time Compute”. We found cases where longer reasoning leads to lower accuracy. Our findings suggest that naïve scaling of test-time compute may inadvertently reinforce problematic reasoning patterns. 🧵
2
1
6