TOPIC: DEEP LEARNING
An AI email newsletter roundup: Cutting through the noise
23rd August 2025This time last year, I felt out of the loop on all things AI. That was put to rights during the autumn when I experimented a lot with GenAI while enhancing travel content on another portal. In addition, I subscribed to enough email newsletters that I feel the need to cull them at this point. Maybe I should use a service like Kill the Newsletter to consolidate things into an RSS feed instead; that sounds like an interesting option for dealing with any overload.
So much is happening in this area that it is too easy to feel overwhelmed by what is happening. That sense got me compiling the state of things in a previous post using some help from GenAI, though I was making the decisions about what was being consolidated and how it was being done. The whole process took a few hours, an effort clearly beyond a single button push.
This survey is somewhat eclectic in its scope; two of the newsletters are hefty items, while others include brevity as part of their offer. Regarding the latter, I found strident criticism of some of them (The Rundown and Superhuman are two that are mentioned) in an article published in the Financial Times, which is behind a paywall. Their content has been called slop, with the phrase slopaganda being coined and used to describe this. That cannot be applied everywhere, though. Any brevity cannot cloak differences in tone and content choices can help with developing a more rounded view of what is going on with AI.
This newsletter came to my notice because I attended SAS Innovate on Tour 2025 in London last June. Oliver Patel, who authors this and serves as Enterprise AI Governance Lead at AstraZeneca as well as contributing to various international organisations including the OECD Expert Group on AI Risk and Accountability, was a speaker with the theme of his talk naturally being AI governance as well as participating in an earlier panel on the day. Unsurprisingly, the newsletter also got a mention.
It provides in-depth practical guidance on artificial intelligence governance and risk management for professionals working in enterprise environments, though not without a focus on scaling governance frameworks across organisations. Actionable insights are emphasised in place of theoretical concepts, covering areas such as governance maturity models that progress from nascent stages through to transformative governance, implementation strategies and leadership approaches needed to drive effective AI governance within companies.
Patel brings experience from roles spanning policy work, academia and privacy sectors, including positions with the UK government and University College London, which informs his practical approach to helping organisations develop robust AI governance structures. The newsletter targets AI governance professionals, risk managers and executives who need clear, scalable solutions for real-world implementation challenges, and all content remains freely accessible to subscribers.
Unlike other newsletters featured here, this is a seven-day publication that delivers a five‑minute digest on AI industry happenings each day that combines news, productivity tips, polls and AI‑generated art. It was launched in June 2023 by Matt Village and Adam Biddlecombe, using of beehiiv’s content‑focused platform that was acquired by HubSpot in March 2025, placing it within the HubSpot Media Network.
Created by Zain Kahn and based in Toronto, weekday issues of this newsletter typically follow a structured format featuring three AI tools for productivity enhancement, two significant AI developments and one quick tutorial to develop practical skills. On Saturdays, there is a round-up on what is happening in robotics, while the Sunday issue centres on developments in science. Everything is crafted to be brief, possibly allowing a three-minute survey of latest developments.
The Artificially Intelligent Enterprise
My interest in the world of DevOps led me to find out about Mark Hinkle, the solopreneur behind Peripety Labs and his in-depth weekly newsletter published every Friday that features comprehensive deep dives into strategic trends and emerging technologies. This has been complemented by a shorter how-to version which focusses on concrete AI lessons and implementation tips and comes out every Tuesday, taking forward a newsletter acquired from elsewhere. The idea is that we should concentrate on concrete AI lessons and implementation tips in place of hype, particularly in business settings. These forms part of The AIE Network alongside complementary publications including AI Tangle, AI CIO and AI Marketing Advantage.
Found though my following the Artificially Intelligent Enterprise, this daily newsletter delivers artificial intelligence developments and insights within approximately five minutes of reading time per issue. Published by Rowan Cheung, it covers key AI developments, practical guides and tool recommendations, with some articles spanning technology and robotics categories. Beyond the core newsletter, the platform operates AI University, which provides certificate courses, implementation guides, expert-led workshops and community networking opportunities for early adopters.
Claude Projects: Reusing your favourite AI prompts
28th March 2025Some things that I do with Anthropic Claude, I end up repeating. Generating titles for pieces of text or rewriting text to make it read better are activities that happen a lot. Others would include the generation of single word previews for a piece or creating a summary.
Python or R scripts come in handy for summarisation, either for a social media post or for introduction into other content. In fact, this is how I go much of the time. Nevertheless, I found another option: using Projects in the Claude web interface.
These allow you to store a prompt that you reuse a lot in the Project Knowledge panel. Otherwise, you need to supply a title and a description too. Once completed, you just add your text in there for the AI to do the rest. Title generation and text rewriting already are set up like this, and keywords could follow. It is a great way to reuse and refine prompts that you use a lot.
OWASP Top 10 for Large Language Model Applications
21st January 2024OWASP stands for Open Web Application Security Project, and it is an online community dedicated to web application security. They are well known for their Top 10 Web Application Security Risks and late last year, they added a Top 10 for
Large Language Model (LLM) Applications.
Given that large language models made quite a splash last year, this was not before time. ChatGPT gained a lot of attention (OpenAI also has had DALL-E for generation of images for quite a while now), there are many others with Anthropic Claude and Perplexity also being mentioned more widely.
Figuring out what to do with any of these is not as easy as one might think. For someone more used to working with computer code, using natural language requests is quite a shift when you no longer have documentation that tells what can and what cannot be done. It is little wonder that prompt engineering has emerged as a way to deal with this.
Others have been plugging in LLM capability into chatbots and other applications, so security concerns have come to light, so far, I have not heard anything about a major security incident, but some are thinking already about how to deal with AI-suggested code that others already are using more and more.
Given all that, here is OWASP's summary of their Top 10 for LLM Applications. This is a subject that is sure to draw more and more interest with the increasing presence of artificial intelligence in our everyday working and no-working lives.
LLM01: Prompt Injection
This manipulates an LLM through crafty inputs, causing unintended actions by the LLM. Direct injections overwrite system prompts, while indirect ones manipulate inputs from external sources.
LLM02: Insecure Output Handling
This vulnerability occurs when an LLM output is accepted without scrutiny, exposing backend systems. Misuse may lead to severe consequences such as Cross-Site Scripting (XSS), Cross-Site Request Forgery (CSRF), Server-Side Request Forgery (SSRF), privilege escalation, or remote code execution.
LLM03: Training Data Poisoning
This occurs when LLM training data are tampered, introducing vulnerabilities or biases that compromise security, effectiveness, or ethical behaviour. Sources include Common Crawl, WebText, OpenWebText and books.
LLM04: Model Denial of Service
Attackers cause resource-heavy operations on LLMs, leading to service degradation or high costs. The vulnerability is magnified due to the resource-intensive nature of LLMs and the unpredictability of user inputs.
LLM05: Supply Chain Vulnerabilities
LLM application lifecycle can be compromised by vulnerable components or services, leading to security attacks. Using third-party datasets, pre-trained models, and plugins can add vulnerabilities.
LLM06: Sensitive Information Disclosure
LLMs may inadvertently reveal confidential data in its responses, leading to unauthorized data access, privacy violations, and security breaches. It’s crucial to implement data sanitization and strict user policies to mitigate this.
LLM07: Insecure Plugin Design
LLM plugins can have insecure inputs and insufficient access control. This lack of application control makes them easier to exploit and can result in consequences such as remote code execution.
LLM08: Excessive Agency
LLM-based systems may undertake actions leading to unintended consequences. The issue arises from excessive functionality, permissions, or autonomy granted to the LLM-based systems.
LLM09: Overreliance
Systems or people overly depending on LLMs without oversight may face misinformation, miscommunication, legal issues, and security vulnerabilities due to incorrect or inappropriate content generated by LLMs.
LLM10: Model Theft
This involves unauthorized access, copying, or exfiltration of proprietary LLM models. The impact includes economic losses, compromised competitive advantage, and potential access to sensitive information.