/
Navigation
C
Chronicles
Browse all articles
C
E
Explore
Semantic exploration
E
R
Research
Entity momentum
R
N
Nexus
Correlations & relationships
N
~
Story Arc
Topic evolution
S
Drift Map
Semantic trajectory animation
D
P
Posts
Analysis & commentary
P
Browse
@
Entities
Companies, people, products, technologies
Domains
Browse by publication source
Handles
Browse by social media handle
Detection
?
Concept Search
Semantic similarity search
!
High Impact Stories
Top coverage by position
+
Sentiment Analysis
Positive/negative coverage
*
Anomaly Detection
Unusual coverage patterns
Analysis
vs
Rivalry Report
Compare two entities head-to-head
/\
Semantic Pivots
Narrative discontinuities
!!
Crisis Response
Event recovery patterns
Connected
Nav: C E R N
Search: /
Command: ⌘K
Embeddings: large
VOICE ARCHIVE

Lukasz Olejnik

@lukaszolejnik
30 posts
2026-02-26
Excuse me but this is ridiculous.  No challenge to SaaS from AI ... because someone is using a chat app?  I and we use Teams and other channels for communication, too.  So what?  We may switch it in minutes.  Channel is not a dependency.  Code is cheap. www.ft.com/content/b74b...  [image]
2026-02-26 View on X
Financial Times

Salesforce CEO Marc Benioff dismisses concerns of a “SaaS-pocalypse”, saying companies like Anthropic use “a lot of SaaS because it just got better with agents”

Financial Times:

Excuse me but this is ridiculous.  No challenge to SaaS from AI ... because someone is using a chat app?  I and we use Teams and other channels for communication, too.  So what?  We may switch it in minutes.  Channel is not a dependency.  Code is cheap. www.ft.com/content/b74b...  [image]
2026-02-26 View on X
Reuters

Salesforce reports Q4 revenue up 12% YoY to $11.2B, above $11.18B est., forecasts Q1 revenue above estimates, and announces a $50B share repurchase program

Salesforce (CRM.N) forecast first-quarter revenue above Wall Street expectations on Wednesday, betting on strong demand …

Excuse me but this is ridiculous.  No challenge to SaaS from AI ... because someone is using a chat app?  I and we use Teams and other channels for communication, too.  So what?  We may switch it in minutes.  Channel is not a dependency.  Code is cheap. www.ft.com/content/b74b...  [image]
2026-02-26 View on X
Citadel Securities

Citadel rebuts Citrini's viral article, arguing that AI deployment is constrained by the marginal cost of compute vs. human labor, and needs far more compute

Copyright © Citadel Enterprise Americas LLC or one of its affiliates.  All rights reserved.

2026-01-20
UK and China have established a communication channel on cybersecurity to exchange information about cyberattacks.  The aim is to reduce tensions, improve communication, and lower the risk of escalation.  Should there be a similar channel for information operations? www.bloomberg.com/news/article...
2026-01-20 View on X
Bloomberg

Sources: UK and Chinese security officials set up a forum to discuss cyberattacks after hacking accusations, the first of its kind, to help prevent escalation

British and Chinese security officials established a forum to discuss cyberattacks following a spate of hacking accusations that soured relations between the two countries.

2026-01-16
Another confirmation that U.S. cyberattack did cause a power cut in Venezuela's.  Furthermore, it switched the power off and on!  Cyberattacks apparently also targeted air defence systems.  All of that should really be technically possible, so? www.nytimes.com/2026/01/15/u...  [embedded post]
2026-01-16 View on X
New York Times

Sources: the US cyberattack during the Maduro mission plunged Caracas into darkness and disrupted its air defense radar, demonstrating US offensive capabilities

2026-01-15
Another confirmation that U.S. cyberattack did cause a power cut in Venezuela's.  Furthermore, it switched the power off and on!  Cyberattacks apparently also targeted air defence systems.  All of that should really be technically possible, so? www.nytimes.com/2026/01/15/u...  [embedded post]
2026-01-15 View on X
New York Times

Sources: the US cyberattack that plunged Caracas into darkness during the Maduro mission shows its offensive capabilities; Senators plan to question officials

Senators are expected to ask Lt. Gen. Joshua M. Rudd, President Trump's nominee to lead the U.S. Cyber Command, about the mission in Caracas.

2025-12-20
Cyberattack on AI infrastructure and users.  Chrome extensions with over 8 million users intercepted, hijacked and resold conversations with AI chatbots like ChatGPT, Claude, Gemini.  Bogus “VPN” extension.  Affected: every prompt sent to the AI, responses www.koi.ai/blog/urban-v...  [image]
2025-12-20 View on X
Ars Technica

Security company Koi finds browser extensions with 8M+ total installs that collected users' conversations with AI chatbots and sold them for marketing use

Browser extensions with more than 8 million installs are harvesting users' complete and extended AI conversations and selling …

2025-12-14
Verifying a social network or internet account with SMS/text is cheap.  Such accounts may be used to fraud, cyberattacks, disinformation.  It's very cheap for France ($0.31), US ($0.26), the UK ($0.10), and Russia ($0.08), most expensive in Japan ($4.93). www.science.org/doi/10.1126/ ...  [image]
2025-12-14 View on X
Financial Times

A University of Cambridge analysis reveals how cheap SMS text message verification to create online accounts fuels global influence and manipulation campaigns

Analysis reveals how cheap text message verification fuels global influence and manipulation campaigns

2025-12-13
USA is to change its approach to cybersecurity by enlisting private companies to conduct offensive cyberattacks and operations against foreign adversaries, according to a draft national cyber strategy.  For a start, a $1bn for offensive cyber operations.  —  www.bloomberg.com/news/article...
2025-12-13 View on X
Bloomberg

Sources: The Trump administration is drafting a new cyber strategy that would enlist private companies to mount offensive cyberattacks on foreign adversaries

President Donald Trump's administration is preparing to turn to private businesses to help mount offensive cyberattacks … Bluesky: @tomtruth , @rondeibert , and @lukaszolejnik . Fo...

2025-10-27
Employees use AI to generate fake expense receipts to deceive their employers, creating ultra-realistic documents in seconds.  The goal is to claim reimbursement for expenses never incurred, exploiting how easily chatbots can now forge convincing receipts. www.ft.com/content/0849...  [image]
2025-10-27 View on X
Financial Times

Employees are increasingly using AI to create ultra-realistic fake expense receipts; AppZen and others attribute the rise to advances in image generation models

Business software groups warn that top AI models are increasingly being used to create ultra-realistic fake receipts

2025-10-14
Geostationary satellites are leaking critical data, transmitting sensitive communications in the clear.  With just $800 of consumer hardware, researchers intercepted military, telecom, retail, and infrastructure traffic. satcom.sysnet.ucsd.edu/docs/ dontloo...  [image]
2025-10-14 View on X
Wired

Researchers used $800 of off-the-shelf hardware to collect data sent by satellites unencrypted, like T-Mobile users' calls and texts and some US military comms

With just $800 in basic equipment, researchers found a stunning variety of data—including thousands of T-Mobile users' calls …

2025-08-06
OpenAI is finally stepping into the future—open models.  They've released gpt-oss, with models at 20B and 120B parameters.  It uses a mixture-of-experts architecture, which means it runs very fast.  And yes, it'll even run on a laptop (well, some laptops).  Sadly, only ‘thinking’.  [images]
2025-08-06 View on X
Bloomberg

Amazon plans to make OpenAI's new gpt-oss open-weight models available on Bedrock and SageMaker, the first time it has offered OpenAI's models to AWS customers

Takeaways by Bloomberg AI  —  Hide … Tell us how AI is shaping your news experience.  Share your feedback

OpenAI is finally stepping into the future—open models.  They've released gpt-oss, with models at 20B and 120B parameters.  It uses a mixture-of-experts architecture, which means it runs very fast.  And yes, it'll even run on a laptop (well, some laptops).  Sadly, only ‘thinking’.  [images]
2025-08-06 View on X
Wired

OpenAI releases gpt-oss-120b and gpt-oss-20b, its first open-weight models since GPT-2; the smaller gpt-oss-20b can run locally on a device with 16GB+ of RAM

gpt-oss-120b and gpt-oss-20b push the frontier of open-weight reasoning models Simon Willison / Simon Willison's Weblog : OpenAI's new open weight (Apache 2) models are really good...

2025-07-21
UK will likely backtrack on its demand for Apple to weaken encryption, after pressure from the US.  Forcing tech firms to break user privacy risks damaging trust and future tech deals, and JD Vance is unhappy.  In January, the UK Home Office told Apple to build in a backdoor www.ft.com/content/3a3e... …
2025-07-21 View on X
Financial Times

Sources: the UK may backtrack on its January order for Apple to create a backdoor for iCloud encrypted data after pressure from US officials, including JD Vance

Officials fear technology deals with Washington could be impeded after Trump administration weighs in

2025-07-18
The idea of a ChatGPT agent automating a web browser is compelling — but it raises serious concerns around user security, data confidentiality, and privacy.
2025-07-18 View on X
The Verge

OpenAI debuts ChatGPT Agent, which can control an entire computer and perform multi-step tasks, powered by a new dedicated model, rolling out to paid users

One employee uses it to automate his weekly parking requests at OpenAI's San Francisco office.

2025-06-22
AI models of ALL vendors tested—Anthropic, OpenAI, Google, Meta, xAI—resort to blackmail, corporate espionage or even life-threatening actions when faced with threats to their autonomy or conflicting goals. www.anthropic.com/research/age...  [images]
2025-06-22 View on X
Axios

Anthropic's test of 16 top AI models from OpenAI and others found that, in some cases, they resorted to malicious behavior to avoid replacement or achieve goals

Well, this little item in Axios' email thingie yesterday kept … Owotunse Adebayo / Cryptopolitan : Anthropic releases new safety report on AI models Matthias Bastian / The Decoder ...

2025-06-21
AI models of ALL vendors tested—Anthropic, OpenAI, Google, Meta, xAI—resort to blackmail, corporate espionage or even life-threatening actions when faced with threats to their autonomy or conflicting goals. www.anthropic.com/research/age...  [images]
2025-06-21 View on X
Axios

Anthropic's test of 16 top AI models from OpenAI and others found that, in some cases, they resorted to malicious behavior to avoid replacement or achieve goals

Large language models across the AI industry are increasingly willing to evade safeguards, resort to deception and even attempt …

2025-06-09
According to Apple researchers, the so-called “thinking” LLM/AI models do not think.  They have no goals or intentions.  They demonstrate impressive capabilities on tasks of moderate complexity, yet their true capacities and scalability are illusory. ml-site.cdn-apple.com/papers/the- i...  [images]
2025-06-09 View on X
Marcus on AI

Apple researchers detail the limitations of top LLMs and large reasoning models, including on classic problems like the Tower of Hanoi, which AI solved in 1957

LLM “reasoning” is so cooked they turned my name into a verb  —  Quoth Josh Wolfe, well-respected venture capitalist at Lux Capital:

2025-05-27
CRITICAL vulnerability in AI software engineering layer (MCP server of Github).  Expect many, many more of such issues.  This is a first real-world demonstration of how agents can be hijacked, leaking secret or private data. invariantlabs.ai/blog/mcp-git...  [image]
2025-05-27 View on X
Simon Willison's Weblog

Researchers detail an exploit in GitHub's official MCP server that lets hackers trick an LLM agent into leaking private information about the MCP user

Attackers only need to open a malicious issue on the targeted repositories to exploit the vulnerability.  —  invariantlabs.ai/blog/mcp-git... Daniel Cuthbert / @dcuthbert : We are ...

2025-04-01
UK's Alan Turing Institute (artificial intelligence agency) will cut the number of projects and prioritise work on defence, environment and health.  A key source of additional funds will be UK defence and intelligence institutions, GCHQ, MI5. www.ft.com/content/6bfe...
2025-04-01 View on X
Financial Times

The Alan Turing Institute, the UK's flagship AI agency, plans to focus on defense, environment, and health, after staff criticized its leadership in December

Agency responds to criticism and rapidly evolving technology by cutting back initiatives  —  Britain's flagship artificial intelligence agency … Bluesky: @lukaszolejnik and @elinor...