TechLetters #178 LLM use to find vulnerabilities (this time seriously). Kaspersky sanctions. AI exposure of employees.
Security
LLM-assisted vulnerability search. Google Project Zero tested it. "Current LLMs can start to perform (admittedly rather basic) vulnerability research ... there's a large difference between solving isolated CTF-style challenges and performing autonomous offensive security research". So disregard the flashy headlines. LLMs have difficulties with resolving ambiguity, and that’s the real world.
USA includes Kaspersky executives on sanctions list. "Department of Commerce has designated AO Kaspersky Lab and OOO Kaspersky Group (Russia), and Kaspersky Labs Limited (United Kingdom) on the Entity List for their cooperation with Russian military". Kaspersky Labs isn’t happy.
Technology Policy
Higher earners are more exposed to the impact of AI than lower paid workers. 20% of employees would see at least half their tasks affected. Occupations with highest exposure: software engineers, data scientists, interpreters and translators, accountants, graphic designers, PR specialists, financial quantitative analysts, etc. Occupations with no exposed tasks: motorcycle mechanics, pile driver operators, dredge operators, fish cutters, tapers, stonemasons.
In case you feel it's worth it to forward this content further:
Subscribed
If you’d like to share: