Researcher and a best-selling author. Keynote talks at RSA, Black Hat & DEF CON. TED Speaker. Chief Research Officer at Sensofusion.

Finland
@mikko retweeted
Backwards compatibility of Windows is amazing. The oldest GUI app I have is *20 YEARS* old today and it works fine under 64-bit Windows 10.
9
56
151
The headquarters of one of the world's most valuable companies was hit by a missile last night.
Reuters: A company source told Reuters earlier that a ballistic missile hit the Aramco headquarters complex in Dhahran. The source said an office tower in the complex was struck and that its top three floors were damaged. The source and two other people with knowledge of the matter said the building was on fire. The complex hosts a major operations centre for Aramco, which oversees a vast network of onshore and offshore oil rigs, pipelines, refineries and export terminals.
7
18
178
19,960
Sensofusion’s IPO has now been successfully completed.   Thank you to all new shareholders for your trust and for joining us in supporting Sensofusion’s mission to protect humans against autonomous weapons.
8
17
379
12,600
@mikko retweeted
Wow, I didn't realize OpenAI was doing math without permission
Important statement from the Association for Human Mathematics ahmath.org/statements
7
10
185
7,853
@mikko retweeted
I don’t think there is any strong argument against today being the singular most important day in the history of mathematics
72
171
2,948
233,994
Welcome to my TED Talk. TED AI at the imperial palace in Vienna, October 28-30, 2026. tedai-vienna.ted.com/speaker…
10
28
307
12,395
@mikko retweeted
💻✨"Rara vez se agradece a alguien por el trabajo que hizo para prevenir el desastre que no sucedió." – Mikko Hypponen 💻✨ #BitsDeInspiración #Programación #Desarrolladores
1
2
6
3,258
Me Sensofusionilla rakennamme järjestelmiä droonien tunnistamiseen ja torjuntaan. Nyt suunnittelemme listautumista Helsingin pörssiin. Ilmoittaudu 1.10. pidettävään webcastiin ja virtuaaliselle tehdaskierrokselle osoitteessa ipo.sensofusion.com kuullaksesi lisää.
17
37
502
28,297
If you find two ants in your kitchen, the best estimate of the total number of ants in your kitchen is not two
We're detailing two new incidents that occurred during external cyber evaluations conducted by independent evaluation partners. We outline what happened, how the activity was contained, and how we’re working with evaluators to strengthen our approach to third-party testing. openai.com/index/third-party…
30
369
5,622
327,131
@mikko retweeted
The incident timeline is wild. It took the monitoring system 12 minutes to notice the agent gained unauthorized internet access and then 2 minutes later a human acknowledged that. And then it took them TWO AND A HALF HOURS to stop the run.
Some new misalignment disclosures from OpenAI: • Last Sunday morning, one of our models was able to gain unauthorized access to the internet during RL training (~all inference for our most capable models remains stopped until we have hardened our systems further) • In May, a version of HPIM uploaded a employee's GitHub token to the internet, causing the model to be quarantined for two weeks • A new research finding, demonstrating that one can construct self-replicating prompt injections alignment.openai.com/misalig…
21
58
748
49,434
@mikko retweeted
“Feel,” even “want,” I can get behind as things for which evidence is thin. But AIs *clearly* think and understand. If your definition of “think” excludes entities that can autonomously resolve Millennium Problems, it is your definition that is flawed. This desire to avoid anthropomorphic language has become a pathology.
Artificial intelligence systems do not think, feel, want or understand. Avoid language that gives them human characteristics. This is called anthropomorphizing, when we ascribe human traits, emotions or behaviors to non-human things, such as animals or inanimate objects. Instead, explain what a system does, how well it performs, who built it and who could be affected by it. apnews.com/article/openai-sa…
166
94
1,180
121,717
Last week, a rogue OpenAI agent probed a cryptocurrency trading platform called Quidax. It repeatedly tried to trade crypto but failed to submit the trades, then tried an HTML injection, and probed the API. transluce.org/agent-activity
5
21
75
12,501
While the agent apparently failed to make money, the attempt in itself is worrying. If a rogue AI agent is scary, a wealthy rogue AI agent is much scarier.
5
4
26
3,140
[encrypted in Finnish] Sensofusion suunnittelee listautumista Helsingin pörssiin. Tervetuloa 1.10.2026 klo 18 yhtiön järjestämään webcastiin ja vetämälleni virtuaaliselle tehdaskierrokselle. Kaikki suunniteltuun listautumiseen liittyvä täältä: ipo.sensofusion.com
13
28
435
52,601
You could fit about 400 million cards like this in a freight container. That container could store roughly one year of all internet traffic.
SanDisk outs the 'world's first' 1TB SD card engt.co/2cOhY4O
13
136
208
The largest cards on sale in 2026 are 2 TB microSD cards. 400 million cards would still physically fit in one 40-foot shipping container, but they could only store about 40 days of today's global internet traffic. Also, they'd weigh over 100 tonnes. The cloud is heavy.
2
1
25
1,882
RT @EugeneDavydov: As we all patiently await the next issue of the legendary @Phrack Cybersecurity publication, this preview article by @mi…
2
570
I'm proud of this one.
Your weekend reading assignment has arrived early. A first taste of the upcoming, still-under-wraps Phrack 73: “THE PROXY THAT MADE NO SENSE” by @mikko. archives.phrack.org/dl/73/th…
4
7
75
14,619
@mikko retweeted
SITUATION DETECTED: Anthropic has set up a wet lab in the San Francisco Bay Area for physical biology work. It wants to unlock treatments for rare diseases, and its research has now gone beyond in silico evaluations, per Reuters.
258
226
3,939
2,024,589
@mikko retweeted
We might be the last generation that writes software by hand. @mikko Full ep: piped.video/watch?v=PQoBJ4lL…
2
6
6,732