v.d. direktorja @URSIV_Slovenia | kibernetska odpornost, varnost v digitalnem okolju.
Republic of Slovenia
Joined September 2011
- Tweets2.4K
- Following861
- Followers4.1K
- Likes6.6K
Glede predstavitve rešitve Sphere podjetja Dream Security želim pojasniti ključna dejstva.
Sestanek 13. julija 2026 je bil predstavitvene narave. URSIV rešitve Sphere doslej ni testiral.
Možnost laboratorijskega preizkusa je bila obravnavana, vendar termin ni določen.
Predstavitev in možnost preizkusa ne predstavljata odločitve o nakupu ali uvedbi rešitve v operativno okolje.
Celotno pojasnilo URSIV: gov.si/novice/2026-10-29-poj…
Umetna inteligenca spreminja hitrost kibernetskih napadov. Čas je, da enako hitro začnemo razvijati tudi obrambo.
Dogodki zadnjih dni kažejo, kako hitro se spreminja področje kibernetske varnosti.
Microsoft poroča o napadalcih, ki z uporabo umetne inteligence v posameznih primerih pridobijo privilegiran dostop do oblačne infrastrukture v manj kot treh urah. CrowdStrike je dokumentiral štiri skupine, ki uporabljajo avtonomne agente, pri čemer je ena kompromitirala najmanj 15 zdravstvenih organizacij.
V drugi dokumentirani kampanji so napadalci z uporabo agentov umetne inteligence poskušali vdreti v več kot 1.200 spletnih mest, raziskovalci pa poročajo o več kot 100 uspešnih vdorih.
Hkrati smo priča kritičnim ranljivostim v široko uporabljeni programski opremi, napadom na programsko dobavno verigo in novim tveganjem, povezanim z avtonomnimi sistemi AI.
Umetna inteligenca razkriva naš varnostni dolg hitreje, kot ga odpravljamo.
Ranljivosti, ki so lahko leta ostajale neodkrite, postajajo vse lažje dostopne. Naloge, za katere so napadalci nekoč potrebovali dneve ali tedne, lahko danes v določenih primerih avtomatizirajo.
To pa ni samo slaba novica.
Ista tehnologija, ki napadalcem pomaga odkrivati ranljivosti, lahko obrambni strani pomaga, da jih odkrije in odpravi še pred napadalci.
Toda to zahteva spremembo načina razmišljanja.
Ni dovolj, da umetno inteligenco uporabljamo za pripravo poročil in povzetkov. Uporabiti jo moramo tudi za neprekinjeno preverjanje izpostavljenosti sistemov, odkrivanje ranljivosti, analizo groženj, pripravo varnostnih popravkov in pospeševanje odzivanja na incidente.
Seveda ob ustreznem nadzoru, preverjanju in jasno določenih odgovornostih.
Naš cilj ne sme biti zgolj slediti hitrosti napadalcev. Naš cilj mora biti, da jih prehitimo.
Kibernetska odpornost v dobi umetne inteligence ne bo odvisna samo od tega, kako dobro znamo preprečevati napade, temveč tudi od tega, kako hitro znamo odkrivati, odpravljati in omejevati ranljivosti.
Hitrost odzivanja postaja eden ključnih elementov nacionalne kibernetske varnosti.
Umetna inteligenca temeljito spreminja ekonomiko kibernetske varnosti. Odkrivanje in izkoriščanje ranljivosti postaja hitrejše, cenejše in dostopnejše.
Ključno vprašanje ni več le, kako preprečiti napade, temveč kako z uporabo umetne inteligence ranljivosti odkriti in odpraviti hitreje, kot jih lahko izkoristijo napadalci.
Prihodnost kibernetske varnosti bo vse bolj tekmovanje v hitrosti prilagajanja.
Some people say this will be the end of PHP
I don't think so because these AI-assisted vulnerability exploits will now increasingly happen to any language, any framework, any server software etc.
I mean Node patched a high security vulnerability this week too
A new era of infinite cat and mouse games between people assisted by super intelligent AI trying to hack every single app, site and server and on the other side companies trying to defend themselves from it also assisted by AI
A new era of cybersecurity has begun
AI models are absorbing the logic once embedded in specialized software.
The next shift isn’t just replacing SaaS. It’s replacing application-centric computing with a cognitive layer capable of dynamically composing and executing arbitrary software operations.
From apps to intent.
Data brokers are now buying up SaaS data sets for anywhere from $100K to $2M
SaaS are struggling now and some extra money to fill the gap by declining revenues is nice
That data is then re-sold to AI training companies and AI frontier labs for the exact purpose of integrating whatever that SaaS did into the AI frontier lab's app
Pretty obvious to see where this is going
Niko Gamulin retweeted
Dusting off this tweet from April because this gap in shared understanding of LLM capability is *widening*. It's now less "two groups of people speaking past each other" and more a sharp funnel.
- Napkin math somewhere around 6B people (~75% of the population) have barely come in contact with LLMs at all.
- Around 1-2B (~20%) are casual and infrequent users of free-tier ChatGPT-like products. This group sees derpy chatbots and treats them a bit like a better Google search, a writing aid, or etc. Maybe an agent tries to book you a flight. I have non-tech friends who (reasonably, imo) say they have not much use for it at all. Even many professionals outside of math&code are in this tier. For example, execs and many other functions spend a lot of their time talking to other people, so while they understand what is happening intellectually, it is still second-hand and a bit abstract.
- Now we get to professional use of frontier-grade LLMs in math&code. Somewhere around 20M people (0.2%) see first-hand that large, complex projects that used to take them weeks/months can now be completed by agents with a prompt. Building apps, copying apps, translating apps, decompiling apps from binaries... This has all happened very quickly and recently - less than 1 year ago, I was writing code manually by hand, typing memorized computer code commands into a code editor character by character, occasionally pressing Tab to autocomplete a little chunk of code.
- And finally we get to the ~5,000 people (~0.00006%) with access to frontier-grade systems internally. The external world has seen the preview. It looks like swarms of thousands of agents collaborating over weeks on software mega projects: minting zero days, running cyber attacks and defenses at machine speeds, discovering new science, advancing the frontier of mathematics. Things that would have taken top professionals in the industry years of work. Meanwhile, human review and comprehension are starting to fall behind. For example, people are still involved in the "archeology" of the OpenAI-HF incident from many months ago. Mathematicians may be poring over the 722 manuscripts on frontier mathematics for a while.
The funnel is driven by a combination of factors:
1. The impact scales with ambition, problem size, and horizon. A question with a paragraph answer barely stresses the system. You need a reservoir of big, difficult problems that you really care about. This aspect drives the consumer / professional dimension of the funnel.
2. The jaggedness of the system (which I have written about a lot separately). Capability peaks in domains that are digital, verifiable and economically valuable. This is because LLM capability emerges from reinforcement learning on verifiable rewards on a curated environment mixture driven by revenue potential. This aspect primarily drives the area (e.g. math&code) dimension of the funnel.
3. Access. Free-tier, paid-tier, internal.
So this is the weirdness of the moment. The general public has mostly not interacted with these systems. When they have, it looks like a derpy chatbot. The majority of professionals still see only a modest uplift. And a small sliver of professionals are experiencing the vertigo of the curve going vertical. And it is all happening at the same time.
Judging by my tl there is a growing gap in understanding of AI capability.
The first issue I think is around recency and tier of use. I think a lot of people tried the free tier of ChatGPT somewhere last year and allowed it to inform their views on AI a little too much. This is a group of reactions laughing at various quirks of the models, hallucinations, etc. Yes I also saw the viral videos of OpenAI's Advanced Voice mode fumbling simple queries like "should I drive or walk to the carwash". The thing is that these free and old/deprecated models don't reflect the capability in the latest round of state of the art agentic models of this year, especially OpenAI Codex and Claude Code.
But that brings me to the second issue. Even if people paid $200/month to use the state of the art models, a lot of the capabilities are relatively "peaky" in highly technical areas. Typical queries around search, writing, advice, etc. are *not* the domain that has made the most noticeable and dramatic strides in capability. Partly, this is due to the technical details of reinforcement learning and its use of verifiable rewards. But partly, it's also because these use cases are not sufficiently prioritized by the companies in their hillclimbing because they don't lead to as much $$$ value. The goldmines are elsewhere, and the focus comes along.
So that brings me to the second group of people, who *both* 1) pay for and use the state of the art frontier agentic models (OpenAI Codex / Claude Code) and 2) do so professionally in technical domains like programming, math and research. This group of people is subject to the highest amount of "AI Psychosis" because the recent improvements in these domains as of this year have been nothing short of staggering. When you hand a computer terminal to one of these models, you can now watch them melt programming problems that you'd normally expect to take days/weeks of work. It's this second group of people that assigns a much greater gravity to the capabilities, their slope, and various cyber-related repercussions.
TLDR the people in these two groups are speaking past each other. It really is simultaneously the case that OpenAI's free and I think slightly orphaned (?) "Advanced Voice Mode" will fumble the dumbest questions in your Instagram's reels and *at the same time*, OpenAI's highest-tier and paid Codex model will go off for 1 hour to coherently restructure an entire code base, or find and exploit vulnerabilities in computer systems. This part really works and has made dramatic strides because 2 properties: 1) these domains offer explicit reward functions that are verifiable meaning they are easily amenable to reinforcement learning training (e.g. unit tests passed yes or no, in contrast to writing, which is much harder to explicitly judge), but also 2) they are a lot more valuable in b2b settings, meaning that the biggest fraction of the team is focused on improving them. So here we are.
Niko Gamulin retweeted
As I posted before, I've cancelled almost all of the SaaS subscriptions for my business
The only thing I still pay for as a business:
- AI inference
- Hetzner server hosting
- Cloudflare storage
- Backblaze backups
- Taxes
Now I started cancelling personal subscriptions too and just vibecoding my own versions or using Claude to take an open source project that does the same and then customize it for myself
All I still pay for personally is:
- Claude Code
- YouTube (content)
- Spotify (content)
- Starlink (internet)
- Phone eSIM
- Personal cloud storage
- Electricity and water (utility)
- Taxes
So if we summarize it for both, I only still pay for:
- AI inference
- Content
- Cloud storage & hosting
- Connectivity
- Electricity & water
- Taxes
So to summarize all my money now goes to:
- Content
- Data centers
- Power plants + water
- Governments
I think we'll be facing a tidal wave of cancellations once people find out how much money they can save
It's just how the economy works
the fundamental basic economics misunderstanding that ppl have esp on here is that they think normal ppl need to use models to build their own software to be disrupted. they obviously don't & they obviously won’t.
but even if just 1% of the population can build software, that's enough to flood the entire supply side with millions of new products collapsing the cost to near zero.
you absolutely don't need universal adoption to destroy scarcity. you just need enough producers to overwhelm demand. you can feel this now already. i never pay for software because i just ask claude to point to something like a chrome extension or a mac app & it’s just there. at some point your agents will also auto update & maintain your software automatically too.
Niko Gamulin retweeted
🇪🇺 V okviru #CyberWeek 2026 v Dublinu, ki ga gosti Irska v času predsedovanja Svetu EU, se je v. d. direktorja @URSIV_Slovenia dr. @NikoGamulin udeležil srečanja direktorjev, pristojnih za kibernetsko varnost.
V ospredju razprave: povezava med umetno inteligenco in kibernetsko varnostjo ter njen vpliv na prihodnjo kibernetsko odpornost.
🤝 Srečanje je omogočilo tudi bilateralne pogovore s predstavniki drugih držav.
#CyberSecurity #AI #URSIV
This is the security model we need for the AI era.
Find vulnerabilities in our own systems before adversaries do. Fix them faster. Continuously shrink the window of opportunity.
The best defence increasingly starts by thinking like the attacker.
Greg Brockman of @OpenAI on the new security loop: point every model release at your own systems, find the vulnerabilities, automate it.
"We took 25% of our production engineers and said, 'Sorry, all your projects are on hold. You are now defending. You are now up-leveling our security architecture. You're going to use the models to find all the holes.' And we found a number of serious issues, and we fixed them."
"We found some new problems, but eventually it saturated. We basically have found, to our knowledge, all of the P0s, all of the critical problems that Astra is smart enough to find. And of course, there will be a new model, there will be a new round."
"You want to be in this tight loop of new cyber capability drops, you deploy it against your systems, you find the new holes, and ideally, you've managed to automate this, what we call defense factory. That's what we're building internally."
@gdb @bhorowitz
AI is not creating our security debt. It is exposing it.
That distinction matters.
When AI finds vulnerabilities faster than we are used to, the wrong response is to slow down the technology so the numbers look less alarming.
The vulnerabilities are already there. And malicious actors will not wait for our regulatory comfort.
Europe should aim to be the place that discovers weaknesses first, fixes them first, and builds the tools that make this possible.
The answer to faster discovery is faster defence, not slower innovation.
Niko Gamulin retweeted
In Ljubljana, billboards usually sell influencers.
So we put up different ones.
Four Slovenians most people have never heard of. Each one changed the world from a country of 2 million people. 🧵
IBM je leta 1981 pričakoval približno 242.000 svojih osebnih računalnikov v petih letih. Pozneje jih je v enem mesecu dobavil 250.000. Intuicija je šibka podlaga za dolgoročno tehnološko strategijo.
Današnji dvomi o ekonomski vzdržnosti UI so legitimni, vendar niso dokončna sodba o njeni prihodnosti. Razvoj ni linearen: tehnološki preskoki lahko povečajo gospodarsko vrednost uspešno opravljenega dela na enoto energije.
Toda možnost napredka ne odpravi tveganja skupne odvisnosti. Sto produktov na isti tehnološki osnovi ne pomeni sto neodvisnih stav.
Razvoj je treba podpreti in hkrati graditi sposobnosti, ki ohranijo vrednost tudi, če današnja prevladujoča tehnologija razočara. Razpršitev tveganja zahteva več neodvisnih tehnoloških poti; več projektov na isti osnovi skupne odvisnosti ne odpravi.
AI can increasingly analyze, optimize, draft, and execute.
But institutions still face one task we can’t outsource to machines: getting humans to agree on goals, priorities, responsibilities, and action.
At this point, alignment may be the ultimate institutional bottleneck.
Niko Gamulin retweeted
🎤 Na konferenci Cyber Resilience by Design, ki jo organizira @BritSloChamber, je @NikoGamulin poudaril, da skladnost z #NIS2 ni končni cilj, ampak začetek procesa krepitve kibernetske odpornosti.
Približno 800 slovenskih bistvenih in pomembnih subjektov je že opravilo samoregistracijo. Ključni izzivi ostajajo upravljanje tveganj, varnost dobavnih verig, pripravljenost na incidente in vpliv umetne inteligence na kibernetsko varnost.
🔐 Kibernetska odpornost zahteva aktivno vlogo vodstva, redno preverjanje ukrepov in nenehno prilagajanje novim grožnjam.
#CyberResilience #CyberSecurity #NIS2 #URSIV
Foto: Jaka Tai Gubenšek
Dober test tehnološke investicije:
Kaj ostane, ko infrastruktura čez deset let zastara?
Če ostanejo predvsem stroški zamenjave, smo tehnologijo kupili.
Če ostanejo ljudje, znanje, podjetja in sposobnost zgraditi naslednjo generacijo, smo zgradili tehnološko moč.
Niko Gamulin retweeted
🇸🇮 Slovenija gosti #BlueOLEx26
💬 V. d. direktorja @URSIV_Slovenia dr. @NikoGamulin je poudaril, da EU CyCLONe @enisa_eu omogoča usklajeno sodelovanje držav članic pri obvladovanju resnih kibernetskih kriz ter temelji na zaupanju, jasnih postopkih in izkušnjah, ki jih države gradijo že pred kriznimi razmerami.
🛡️ V ospredju srečanja so izkušnje iz #CyberEurope2026, izboljševanje sodelovanja med tehnično, operativno in politično ravnjo ter prenos ugotovitev iz vaj v konkretne izboljšave.
🤝 Ob robu srečanja se je dr. Niko Gamulin srečal tudi s @ChristianeEU iz Evropske komisije #DGCONNECT.
#URSIV #EUCyCLONe #CyberSecurity
Niko Gamulin retweeted
Bravo 🙌 Slovenija 🇸🇮 gre naprej @elonmusk
⚠️ POZOR: PHISHING
Z ugrabljenega X računa @petra_jansa se še vedno pošiljajo phishing sporočila. Tudi sam sem prejel spodnje sporočilo. Ne klikajte povezave in ne vnašajte svojih podatkov.
Napad je prepričljiv, ker zlorablja identiteto novinarke in zaupanje, ki ga je s svojim delom zgradila.
Kako ga prepoznati?
→ nepričakovano zasebno sporočilo
→ poziv h kliku na zunanjo povezavo
→ domena nima zveze s storitvijo, na katero se sporočilo sklicuje
→ ustvarjen je verjeten razlog za klik
Znano ime ali račun še ne pomeni, da je sporočilo varno. Če je račun ugrabljen, napadalec podeduje tudi njegovo verodostojnost.
Niko Gamulin retweeted
We're publishing our most detailed threat intelligence report to date.
It covers how people tried to misuse Claude—for cyberattacks, influence operations, surveillance, biology, and building weapons—and how we found and stopped them.
We disrupted every operation in the report, and used the lessons from them to strengthen our safeguards. Where appropriate, we also shared what we found with authorities and other AI companies.
These cases are not typical: we’re highlighting some of the most sophisticated misuse we’ve seen. But they’re especially important to discuss, because they show us where AI misuse is headed, where our safeguards work, and where they need to improve.
We’re publishing this report so others can spot the same activity on their own platforms, and so we can give the public a clearer view of how emerging threats develop.
Read the report: anthropic.com/threat-intelli…
Niko Gamulin retweeted
🇸🇮🤝🇭🇺 Ob robu srečanja predstavnikov pristojnih nacionalnih organov za kibernetsko varnost, povezanih v Srednjeevropsko platformo za kibernetsko varnost #CECSP, v Mariboru, se je v. d. direktorja @URSIV_Slovenia dr. @NikoGamulin srečal z direktorjem madžarskega Nacionalnega centra za kibernetsko varnost (#NCSC) Lajosom Szabóm.
Sogovornika sta izmenjala poglede na aktualne kibernetske izzive in potrdila pomen tesnega sodelovanja pri krepitvi kibernetske odpornosti v regiji.
#CyberSecurity #KibernetskaVarnost #Slovenia #Hungary #URSIV
