Koen Simoens
@koen-simoens.bsky.social
Attended a webinar by Ash Maurya yesterday: 25 of 95 participants were AI notetakers! We talk about LLM training cost and waste of “please” in prompts. But multiple tools doing the same thing? Not very efficient 🙃 Maybe the bots need a meeting or we build smarter tools. Got ideas for reducing waste?
May 1, 2025 at 8:17 AM
Attended a webinar by Ash Maurya yesterday: 25 of 95 participants were AI notetakers! We talk about LLM training cost and waste of “please” in prompts. But multiple tools doing the same thing? Not very efficient 🙃 Maybe the bots need a meeting or we build smarter tools. Got ideas for reducing waste?
Reposted by Koen Simoens
JP Morgan's CISO highlights critical #ThirdPartyRisk in SaaS, urging providers to prioritize security over rapid features.
"Secure and resilient by design" is key.
Read the full letter here: www.jpmorgan.com/technology/t...
#CyberSecurity
"Secure and resilient by design" is key.
Read the full letter here: www.jpmorgan.com/technology/t...
#CyberSecurity
An Open Letter to Third-Party Suppliers
From startups to legacy brands, you're making your mark. We're here to help.
www.jpmorgan.com
April 29, 2025 at 8:01 AM
JP Morgan's CISO highlights critical #ThirdPartyRisk in SaaS, urging providers to prioritize security over rapid features.
"Secure and resilient by design" is key.
Read the full letter here: www.jpmorgan.com/technology/t...
#CyberSecurity
"Secure and resilient by design" is key.
Read the full letter here: www.jpmorgan.com/technology/t...
#CyberSecurity
Reposted by Koen Simoens
An open letter to third-party suppliers - JP Morgan - "The modern ‘software as a service’ (SaaS) delivery model is quietly enabling cyber attackers and – as its adoption grows – is creating a substantial vulnerability that is weakening the global economic system"
An open letter to third-party suppliers - JP Morgan - "The modern ‘software as a service’ (SaaS) delivery model is quietly enabling cyber attackers and – as its adoption grows – is creating a substantial vulnerability that is weakening the global economic system"
www.jpmorgan.com
April 28, 2025 at 12:54 PM
An open letter to third-party suppliers - JP Morgan - "The modern ‘software as a service’ (SaaS) delivery model is quietly enabling cyber attackers and – as its adoption grows – is creating a substantial vulnerability that is weakening the global economic system"
Reposted by Koen Simoens
Real-world gen AI use cases from the world's leading organizations | Google Cloud Blog #SuggestedRead #devopsish cloud.google.com/tra...
Real-world gen AI use cases from the world's leading organizations | Google Cloud Blog
Gen AI is everywhere, as top companies, governments, researchers, and startups showcase how they're already using Google's AI solutions to enhance their work.
cloud.google.com
April 18, 2025 at 6:26 PM
Real-world gen AI use cases from the world's leading organizations | Google Cloud Blog #SuggestedRead #devopsish cloud.google.com/tra...
Reposted by Koen Simoens
100 days until the next #EUAIAct milestone: Which rules kick in on 2 August 2025?
Read our latest blog to find out buff.ly/13AcdV5 :
• Notified Bodies
• GPAI Models
• Governance
• Confidentiality
• Penalties
Read our latest blog to find out buff.ly/13AcdV5 :
• Notified Bodies
• GPAI Models
• Governance
• Confidentiality
• Penalties
buff.ly
April 25, 2025 at 7:09 AM
100 days until the next #EUAIAct milestone: Which rules kick in on 2 August 2025?
Read our latest blog to find out buff.ly/13AcdV5 :
• Notified Bodies
• GPAI Models
• Governance
• Confidentiality
• Penalties
Read our latest blog to find out buff.ly/13AcdV5 :
• Notified Bodies
• GPAI Models
• Governance
• Confidentiality
• Penalties
Reposted by Koen Simoens
What is AI Governance and why does it matter?
It’s more than compliance. It’s about setting direction, trust & accountability in AI.
Now also in an accessible and engaging audio format🎧 thanks to NotebookLM.
👉 Read or listen here: buff.ly/iEMQonI
#AIGovernance
It’s more than compliance. It’s about setting direction, trust & accountability in AI.
Now also in an accessible and engaging audio format🎧 thanks to NotebookLM.
👉 Read or listen here: buff.ly/iEMQonI
#AIGovernance
What is AI governance? - The Governors
Exploring what is AI governance: an AI governance definition and frameworks to navigate ethical challenges and ensure responsible AI use.
buff.ly
April 22, 2025 at 10:16 AM
What is AI Governance and why does it matter?
It’s more than compliance. It’s about setting direction, trust & accountability in AI.
Now also in an accessible and engaging audio format🎧 thanks to NotebookLM.
👉 Read or listen here: buff.ly/iEMQonI
#AIGovernance
It’s more than compliance. It’s about setting direction, trust & accountability in AI.
Now also in an accessible and engaging audio format🎧 thanks to NotebookLM.
👉 Read or listen here: buff.ly/iEMQonI
#AIGovernance
Reposted by Koen Simoens
Slopsquatting: when LLM-powered coding assistants make up software dependencies of packages that don't exist.
www.theregister.com/2025/04/12/a...
www.theregister.com/2025/04/12/a...
AI code suggestions sabotage software supply chain
: Hallucinated package names fuel 'slopsquatting'
www.theregister.com
April 16, 2025 at 1:14 PM
Slopsquatting: when LLM-powered coding assistants make up software dependencies of packages that don't exist.
www.theregister.com/2025/04/12/a...
www.theregister.com/2025/04/12/a...
Reposted by Koen Simoens
Hardly the most important issue in government right now, but glad this got fixed at the 11th hour.
Really interested to see what www.thecvefoundation.org does (or doesn't do) in response to the news of new contract funding.
Really interested to see what www.thecvefoundation.org does (or doesn't do) in response to the news of new contract funding.
April 16, 2025 at 1:09 PM
Hardly the most important issue in government right now, but glad this got fixed at the 11th hour.
Really interested to see what www.thecvefoundation.org does (or doesn't do) in response to the news of new contract funding.
Really interested to see what www.thecvefoundation.org does (or doesn't do) in response to the news of new contract funding.
Direct link to the article in The Register www.theregister.com/2025/04/12/a...
April 16, 2025 at 1:11 PM
Direct link to the article in The Register www.theregister.com/2025/04/12/a...
Reposted by Koen Simoens
Slopsquatting
As AI coding assistants invent nonexistent software libraries to download and use, enterprising attackers create and upload libraries with those names -- laced with malware, of course.
As AI coding assistants invent nonexistent software libraries to download and use, enterprising attackers create and upload libraries with those names -- laced with malware, of course.
Slopsquatting
As AI coding assistants invent nonexistent software libraries to download and use, enterprising attackers create and upload libraries with those names -- laced with malware, of course.
www.schneier.com
April 15, 2025 at 4:03 PM
Slopsquatting
As AI coding assistants invent nonexistent software libraries to download and use, enterprising attackers create and upload libraries with those names -- laced with malware, of course.
As AI coding assistants invent nonexistent software libraries to download and use, enterprising attackers create and upload libraries with those names -- laced with malware, of course.
Reposted by Koen Simoens
Practical AI Literacy Training for Startups
Hands-on training aligned with Art. 4 of the EU AI Act. Designed specifically for fast-moving tech companies.
No legalese. No filler. Just what you need to stay compliant and keep shipping.
Details here: buff.ly/ZPbBMqo
#AIforStartups
Hands-on training aligned with Art. 4 of the EU AI Act. Designed specifically for fast-moving tech companies.
No legalese. No filler. Just what you need to stay compliant and keep shipping.
Details here: buff.ly/ZPbBMqo
#AIforStartups
Practical AI Literacy Training for Startups and Scaleups - The Governors
Hands-on, expert-led AI literacy training for startups and scaleups. Aligned with the EU AI Act and designed for fast-moving tech companies.
www.thegovernors.eu
April 15, 2025 at 1:15 PM
Practical AI Literacy Training for Startups
Hands-on training aligned with Art. 4 of the EU AI Act. Designed specifically for fast-moving tech companies.
No legalese. No filler. Just what you need to stay compliant and keep shipping.
Details here: buff.ly/ZPbBMqo
#AIforStartups
Hands-on training aligned with Art. 4 of the EU AI Act. Designed specifically for fast-moving tech companies.
No legalese. No filler. Just what you need to stay compliant and keep shipping.
Details here: buff.ly/ZPbBMqo
#AIforStartups
Reposted by Koen Simoens
Stanford’s 2025 AI Index Report buff.ly/IRIVfT8
🚨 AI incidents ↑ 56.4%, but 89% of orgs report none. Are risks being missed?
🌍 #AIGovernance surged: EU AI Act, UN, OECD + new safety institutes
📉 HELM Safety & AIR-Bench show promise, RAI benchmarks still lag
🚨 AI incidents ↑ 56.4%, but 89% of orgs report none. Are risks being missed?
🌍 #AIGovernance surged: EU AI Act, UN, OECD + new safety institutes
📉 HELM Safety & AIR-Bench show promise, RAI benchmarks still lag
The 2025 AI Index Report | Stanford HAI
Get the latest news, advances in research, policy work, and education program updates from HAI in your inbox weekly.
buff.ly
April 10, 2025 at 3:58 PM
Stanford’s 2025 AI Index Report buff.ly/IRIVfT8
🚨 AI incidents ↑ 56.4%, but 89% of orgs report none. Are risks being missed?
🌍 #AIGovernance surged: EU AI Act, UN, OECD + new safety institutes
📉 HELM Safety & AIR-Bench show promise, RAI benchmarks still lag
🚨 AI incidents ↑ 56.4%, but 89% of orgs report none. Are risks being missed?
🌍 #AIGovernance surged: EU AI Act, UN, OECD + new safety institutes
📉 HELM Safety & AIR-Bench show promise, RAI benchmarks still lag
Reposted by Koen Simoens
The EC unveiled its AI continent action plan to boost AI capabilities. Key areas:
1) Large-scale AI compute infra
2) High-quality data access
3) Strategic sectors
4) Skills and talents
5) Simplify #AIAct implementation, incl. a new AI Act Service Desk
🔗 Read more buff.ly/mb6HOmv
#AIgovernance
1) Large-scale AI compute infra
2) High-quality data access
3) Strategic sectors
4) Skills and talents
5) Simplify #AIAct implementation, incl. a new AI Act Service Desk
🔗 Read more buff.ly/mb6HOmv
#AIgovernance
AI continent
Read about how the EU plans to promote artificial intelligence development in Europe, from building AI infrastructure to strengthening AI skills and talents
commission.europa.eu
April 10, 2025 at 9:21 AM
The EC unveiled its AI continent action plan to boost AI capabilities. Key areas:
1) Large-scale AI compute infra
2) High-quality data access
3) Strategic sectors
4) Skills and talents
5) Simplify #AIAct implementation, incl. a new AI Act Service Desk
🔗 Read more buff.ly/mb6HOmv
#AIgovernance
1) Large-scale AI compute infra
2) High-quality data access
3) Strategic sectors
4) Skills and talents
5) Simplify #AIAct implementation, incl. a new AI Act Service Desk
🔗 Read more buff.ly/mb6HOmv
#AIgovernance
Reposted by Koen Simoens
LLM Reasoning Models don't always say what they think
Excellent article from Anthropic
www.anthropic.com/research/rea...
Excellent article from Anthropic
www.anthropic.com/research/rea...
Reasoning models don't always say what they think
Research from Anthropic on the faithfulness of AI models' Chain-of-Thought
www.anthropic.com
April 3, 2025 at 7:00 PM
LLM Reasoning Models don't always say what they think
Excellent article from Anthropic
www.anthropic.com/research/rea...
Excellent article from Anthropic
www.anthropic.com/research/rea...
Reposted by Koen Simoens
You're experimenting with AI, but not getting real results.
Your teams use AI, but you lack clear policies and accountability.
You know compliance, but AI adds a new level of complexity.
Sounds familiar? Reach out to start your AI governance journey with our 2-week workshop series.
buff.ly/vB2M2hy
Your teams use AI, but you lack clear policies and accountability.
You know compliance, but AI adds a new level of complexity.
Sounds familiar? Reach out to start your AI governance journey with our 2-week workshop series.
buff.ly/vB2M2hy
April 2, 2025 at 2:11 PM
You're experimenting with AI, but not getting real results.
Your teams use AI, but you lack clear policies and accountability.
You know compliance, but AI adds a new level of complexity.
Sounds familiar? Reach out to start your AI governance journey with our 2-week workshop series.
buff.ly/vB2M2hy
Your teams use AI, but you lack clear policies and accountability.
You know compliance, but AI adds a new level of complexity.
Sounds familiar? Reach out to start your AI governance journey with our 2-week workshop series.
buff.ly/vB2M2hy
Reposted by Koen Simoens
New Harvard Business School study on the value of open source software finds that the supply-side value of widely-used OSS is $4.15 billion, but its demand-side value is a massive $8.8 trillion. Without OSS, firms would need to spend 3.5x more on software. www.hbs.edu/ris/Publicat...
www.hbs.edu
March 20, 2025 at 10:16 PM
New Harvard Business School study on the value of open source software finds that the supply-side value of widely-used OSS is $4.15 billion, but its demand-side value is a massive $8.8 trillion. Without OSS, firms would need to spend 3.5x more on software. www.hbs.edu/ris/Publicat...
Reposted by Koen Simoens
i feel compelled to remind everyone that chatgpt/LLMs are functionally the *opposite* of an actual expert.
the more your problems are unusual or novel, the more likely LLMS are to be wrong. they are least helpful when it's most important
the more your problems are unusual or novel, the more likely LLMS are to be wrong. they are least helpful when it's most important
March 19, 2025 at 3:14 PM
i feel compelled to remind everyone that chatgpt/LLMs are functionally the *opposite* of an actual expert.
the more your problems are unusual or novel, the more likely LLMS are to be wrong. they are least helpful when it's most important
the more your problems are unusual or novel, the more likely LLMS are to be wrong. they are least helpful when it's most important
Reposted by Koen Simoens
AI is a climate disaster unless we can find renewable and sustainable energy to power it. ⚡️💚
www.nytimes.com/interactive/...
www.nytimes.com/interactive/...
How A.I. Is Changing the Way the World Builds Computers (Gift Article)
Tech companies are revamping computing — from how tiny chips are built to the way they are arranged, cooled and powered — in the race to build artificial intelligence that recreates the human brain.
www.nytimes.com
March 17, 2025 at 12:21 PM
AI is a climate disaster unless we can find renewable and sustainable energy to power it. ⚡️💚
www.nytimes.com/interactive/...
www.nytimes.com/interactive/...
Computer -- I mean Cursor -- says no.
Cursor AI refused to write code, telling dev: "I cannot generate code for you, as that would be completing your work"
No more vibe coding :-/
I would reinforce this as a guardrail to reduce liability and promote better coding.
arstechnica.com/ai/2025/03/a...
Cursor AI refused to write code, telling dev: "I cannot generate code for you, as that would be completing your work"
No more vibe coding :-/
I would reinforce this as a guardrail to reduce liability and promote better coding.
arstechnica.com/ai/2025/03/a...
AI coding assistant refuses to write code, tells user to learn programming instead
Cursor AI tells user, “I cannot generate code for you, as that would be completing your work.”…
arstechnica.com
March 13, 2025 at 10:23 PM
Computer -- I mean Cursor -- says no.
Cursor AI refused to write code, telling dev: "I cannot generate code for you, as that would be completing your work"
No more vibe coding :-/
I would reinforce this as a guardrail to reduce liability and promote better coding.
arstechnica.com/ai/2025/03/a...
Cursor AI refused to write code, telling dev: "I cannot generate code for you, as that would be completing your work"
No more vibe coding :-/
I would reinforce this as a guardrail to reduce liability and promote better coding.
arstechnica.com/ai/2025/03/a...
The article references Hugging Face's AI Energy Score Leaderboard. Mitsua, the most energy-efficient for image generation (187Wh/image), consumes roughly the same energy as 14 hrs on my laptop (battery). Electricity costs around €0.38/kWh in BE. An image would cost €0.07.
Feels small, but it isn't.
Feels small, but it isn't.
A feature in Nature examines how much energy AI data centers could consume and how researchers want more transparency on those numbers.“We’re just doing our best, trying all kinds of tricks to come up with some kind of numbers.”
How much energy will AI really consume? The good, the bad and the unknown
Researchers want firms to be more transparent about the electricity demands of artificial intelligence.
go.nature.com
March 13, 2025 at 5:20 PM
The article references Hugging Face's AI Energy Score Leaderboard. Mitsua, the most energy-efficient for image generation (187Wh/image), consumes roughly the same energy as 14 hrs on my laptop (battery). Electricity costs around €0.38/kWh in BE. An image would cost €0.07.
Feels small, but it isn't.
Feels small, but it isn't.
Reposted by Koen Simoens
🤔 Does your software qualify as an AI system under the EU AI Act?
In our latest blog post, we look at the European Commission’s guidelines and provide practical questions and examples to help you determine if your system qualifies as an AI system.
🔎 Read the full blog post here: buff.ly/CDljuKy
In our latest blog post, we look at the European Commission’s guidelines and provide practical questions and examples to help you determine if your system qualifies as an AI system.
🔎 Read the full blog post here: buff.ly/CDljuKy
How to decide if your system qualifies as an AI system? - The Governors
In this post we unpack the definition of an AI system under the AI Act. With guiding questions and examples we help you decide if your system is an AI system.
buff.ly
March 13, 2025 at 4:59 PM
🤔 Does your software qualify as an AI system under the EU AI Act?
In our latest blog post, we look at the European Commission’s guidelines and provide practical questions and examples to help you determine if your system qualifies as an AI system.
🔎 Read the full blog post here: buff.ly/CDljuKy
In our latest blog post, we look at the European Commission’s guidelines and provide practical questions and examples to help you determine if your system qualifies as an AI system.
🔎 Read the full blog post here: buff.ly/CDljuKy
Sometimes we overdo it in Belgium: 26 fundamental rights protection authorities 😅
digital-strategy.ec.europa.eu/en/policies/...
Direct link: ec.europa.eu/newsroom/dae...
#AIAct
digital-strategy.ec.europa.eu/en/policies/...
Direct link: ec.europa.eu/newsroom/dae...
#AIAct
ec.europa.eu
March 7, 2025 at 2:30 PM
Sometimes we overdo it in Belgium: 26 fundamental rights protection authorities 😅
digital-strategy.ec.europa.eu/en/policies/...
Direct link: ec.europa.eu/newsroom/dae...
#AIAct
digital-strategy.ec.europa.eu/en/policies/...
Direct link: ec.europa.eu/newsroom/dae...
#AIAct
Picked this up from LinkedIn. A global AI regulation tracker giving you a comprehensive overview of AI regulations, developments and news across the globe, including an #AIGovernance catalogue of frameworks, policies, and templates.
https://buff.ly/4ibf4Hb
Nice work by Raymond Sun.
https://buff.ly/4ibf4Hb
Nice work by Raymond Sun.
Global AI Regulation Tracker
An interactive world map that tracks AI laws, news and resources around the world.
buff.ly
February 27, 2025 at 10:07 AM
Picked this up from LinkedIn. A global AI regulation tracker giving you a comprehensive overview of AI regulations, developments and news across the globe, including an #AIGovernance catalogue of frameworks, policies, and templates.
https://buff.ly/4ibf4Hb
Nice work by Raymond Sun.
https://buff.ly/4ibf4Hb
Nice work by Raymond Sun.