Dogs, AI, and Gyrogears (it’s a slow security news week) – ESW #346

$25M Series B led by Glilot Capital Partners, for a total of $55M in funding. Anecdote's founders previously worked at IntSights prior to its exit to Rapid7. Anecdote is one of dozens of startups trying to easy the pain of GRC and compliance with AI and automation.

Not strictly a security product, Helios acquires application runtime data. Important for debugging and security alike.

ChatGPT Teams makes it possible for dozens of folks to all work within the same ChatGPT datasets and chat histories. Sounds like a small thing, but IMO, collaboration will accelerate AI use cases and refinement within organizations (including criminal ones).

We talked about these layoffs last week, but they didn't make it into the notes, so I'm including them here.

We talked about these layoffs last week, but they didn't make it into the notes, so I'm including them here.

Mike Privette puts together a nice 2023 wrapup - summarizing and analyzing 50+ newsletters he published last year. A few highlights:

2023 at a glance:

• 684 funding rounds across 100+ unique product categories worth ~$12.7B
• 259 M&A transactions across 70+ unique product categories worth ~$40.5B
• 2023 Funding rounds decreased ~12% (from 774 in 2022), and M&A transaction volume dropped ~3% (from 266 in 2022).
• Total 2023 funding dollars dropped ~38% from the ~$20.6B in 2022, and overall M&A decreased ~21% from the ~$51.2B in 2022.
• In short, fewer companies received funding, and those that did got much smaller checks.
• AI was the clear winner in terms of category growth, with 3932% more AI-based funding than 2022.

Check out the comprehensive post for roughly 1000% more insights.

We also chatted with Mike in December to discuss some of these insights: https://youtu.be/RXMWR6IgYjI

An interesting report that focuses on cloud-based adoption of AI from Wiz.

As promised, I'm reporting back on what I learned taking the AI for Personal Productivity course. I learned:

• how to set up LLMs to use custom instructions for everything I ask them
• that some LLMs are significantly better for some tasks than others, and vice versa
• the idea that I'd be using different LLMs for different tasks hadn't occurred to me
• Claude and Perplexity are the two I'm going to be spending a lot more time with
• the professor, Taylor Malmsheimer, suggested making a list of "high brainpower" and "high importance" tasks to assess
• then, we considered these tasks across 3 different AI roles: assistant, strategist, and creator
• she provided tons of examples and suggested prompts for each of these roles
• the examples REALLY helped - there were a ton of tasks I wouldn't have considered throwing at an AI that I actually got some really interesting results from
• analyzing data from surveys was a really compelling example
• there were a ton of great suggestions in the Zoom chat from the ~400 attendees
• one of the attendees even used an LLM to analyze the chat from the workshop, and shared the results before the end of the class!
• she urged us to try to make new habits where we've matched up some of these LLMs with time-saving workflows
• at the end of the day, it's just another tool, but an important one that rates up there with your professional network, google/wikipedia, and research reports
• She does this 2 hour workshop once a month, and it's $250! I don't get any kickbacks or anything if you sign up, but I think it's worth it. I didn't pay anything, since I pay for the annual unlimited membership

A few things in last week's Cyber Why caught my attention:

• Carta, SaaS for cap table management, appears to have crossed the streams, and shut down part of their business as a result
• Startup Unicorns get redefined as $3B+ valuation???

By my count, there are 44 cybersecurity unicorns right now (I use CBInsight's data). If we move the bar up to $3B, that number goes down to 15. The other problem is that NONE of the unicorns on this list have seen a new valuation since valuations took a beating in mid-2022. We saw Cybereason lose an order of magnitude in value, and Snyk (rumored to) lose 50%, so it's anyone's guess as to which companies are or are not a unicorn on this list.

The latest edition of The Reformed Analyst explores a unique FTC ruling. One that bans the Rite-Aid chain of drugstores from using facial recognition technology for the next five years.

I bought one. Why is this such an interesting innovation? Also, how the hell will these new AI gadgets make money, when virtual assistants have been an absolute profit black hole for tech giants?

Incredible that a single source of bad data - a single software system - could lead to so many systematic wrongful arrests.

"...more than 700 sub-postmasters (Post Office franchisees) were wrongfully prosecuted for fraud, false accounting and theft over a 15-year period, with many imprisoned, losing their livelihoods and facing bankruptcy."