Safety
Anthropic Just Put $100M Into AI-Powered Security
Project Glasswing brings together AWS, Apple, Google, Microsoft, NVIDIA, JPMorganChase, CrowdStrike, and others around a single idea: AI has gotten good enough at finding software vulnerabilities that we need to use it defensively — right now — before those same capabilities spread to bad actors.
The centerpiece is Claude Mythos Preview, an unreleased frontier model that has already found thousands of zero-day vulnerabilities across every major operating system and web browser. One example was a 27-year-old flaw that survived decades of human review. Another was found despite over 5 million reviews by automated software.
Anthropic is committing $100M in usage credits and $4M in donations to scan and patch critical infrastructure before attackers get access.
Read more →
Spotlight: Demis Hassabis Is Solving Problems We Thought Were Unsolvable
I typically refrain from providing my opinions on the polarizing figures in AI. Details shift too fast, and taking a stance usually backfires. But if there's one person I'm most intrigued to watch, it's Demis Hassabis.
You've heard of Sam Altman (ChatGPT) and Dario Amodei (Anthropic). Hassabis runs Google DeepMind and Isomorphic Labs — and his work may end up being the most consequential story in AI. Isomorphic's research on protein folding and computational medicine is the most optimistic narrative I follow in this space. It often feels like my newsfeeds are overwhelmed with negativity, but following work that is actually being done to solve real-world problems — disease, cancer — is motivating and exciting.
Both of these are worth your time:
Same question. One extra line. Watch what changes.
Here's something I noticed early on when I started using these tools: I kept getting back walls of text. Paragraphs I didn't ask for. The AI would answer the question, but I'd have to dig through five sentences or five paragraphs to find the one thing I actually needed.
For some people this is fine. For others like me, who want a simpler approach, it turns out that specifying the format can get you a more powerful answer.
One thing to keep in mind as we venture through this journey together: context is everything. Yes, these models have been trained on "all of the world's knowledge" — aka the internet — but they don't have your experiences, and they don't know exactly what you're looking for until you tell them. The trick is knowing how to communicate with them.
When I first started using AI, I heard a word I would never use in my day-to-day — "Tabular." To this day when I hear it, I think of the phrase "Totally tubular." I was a Teenage Mutant Ninja Turtles fan growing up (still am), don't judge me. But why am I telling you this?
Someone showed me what happens if I say "provide this in tabular format" when speaking to AI — and it instantly changed not just how I was building my prompts, but how I was using the models. I was looking to purchase a new TV and was comparing different models. I asked an LLM and it spit out five pages. I then stopped, said give me the price, differences, and similarities in tabular format — and out came a simple table, easy to digest, with my choice staring me in the face. The bigger TV. It also got me thinking:
How else can we dictate what the output is?
- As a table
- In bullet points (one sentence each)
- Checklist
- In 3 sentences or less
Turns out, you have more control than you think.
If you've ever drafted an email using an LLM, you probably already do this almost subconsciously — "this is too long, rewrite it" or "change the tone to be more conversational." The real magic is what happens when you decide how things should look from the beginning. Controlling the context gives you more controlled responses. Check the examples below.
Example 1 — No Format Specified
The Prompt
What are the pros and cons of leasing vs. buying a work truck for a small contracting company?
What the AI gave back ↓
Great question — here's how the two options generally stack up for a small contracting company:
Leasing tends to work well when you want lower monthly payments and predictable costs. You avoid tying up a large chunk of capital upfront, which can matter a lot for a small business managing cash flow. Maintenance is often simpler since the truck is typically under warranty for the lease term, and you can upgrade to a newer vehicle every few years. On the tax side, lease payments are usually deductible as a business expense. The downsides: you don't build equity, there are mileage limits that can be punishing for contractors who drive a lot, and you'll face fees for excess wear and tear — which is almost inevitable on a work truck.
Buying (whether outright or financed) makes more sense when you plan to keep the truck long-term and put heavy miles on it. You own the asset, can customize it however you want, and there are no mileage penalties. Depreciation and loan interest can be deducted, and Section 179 expensing can let you write off a significant portion (sometimes the full cost) in the first year for qualifying vehicles. The trade-offs are higher upfront or monthly costs, you bear all maintenance and repair risk once the warranty expires, and you're stuck with a depreciating asset if your needs change.
A few practical considerations that often tip the decision:
Mileage is usually the biggest factor. If you're running 25,000+ miles a year on job sites, leasing penalties can eat you alive. Cash flow vs. equity is the second — leasing preserves cash, buying builds an asset on your balance sheet. How hard the truck gets used matters too; if it's hauling heavy loads and getting beat up, lease-end charges can be steep. And tax strategy is worth discussing with your accountant, since Section 179 can make buying dramatically more attractive in certain years.
For a lot of small contractors, buying a reliable used truck (2–3 years old, off-lease) and financing it hits a sweet spot — lower acquisition cost, still some useful life left, and full ownership flexibility. But it really depends on your specific situation.
Would you like to dig into any of these angles further — like the tax implications or how to run the numbers on a specific scenario?
It's all there. But you'd never forward this to your business partner and say "here, read this."
Example 2 — With a Format
The Prompt
What are the pros and cons of leasing vs. buying a work truck for a small contracting company? Compare in a table.
What the AI gave back ↓
Here's a comparison for a small contracting company weighing lease vs. buy for a work truck:
| Leasing | Buying |
|---|
| Upfront cost | Low — minimal or no down payment | High — down payment + taxes + fees |
| Monthly payment | Generally lower | Higher (loan) or none (cash purchase) |
| Ownership | No equity built; you return it | You own an asset with resale value |
| Tax treatment | Lease payments typically deductible as a business expense | Depreciation deductions (Section 179 / bonus depreciation can be significant) |
| Mileage / use limits | Mileage caps and excess-wear penalties — tough for heavy field use | No restrictions; drive and use it as hard as you need |
| Customization | Limited — racks, wraps, and tool boxes may violate lease terms | Full freedom to upfit, wrap, and modify |
| Maintenance | Often covered or under warranty for the lease term | Your responsibility, costs rise as truck ages |
| Cash flow | Predictable, fixed monthly expense | Less predictable; big repair bills possible on older trucks |
| End of term | Return it and start over (or buy out) | Keep driving it payment-free for years |
| Turnover / image | Newer truck every 2–4 years | Truck ages unless you sell and replace |
| Best fit | Companies prioritizing low cash outlay, light-duty use, and always-new vehicles | Companies that put on heavy miles, need custom upfits, or want long-term cost savings |
Bottom line for most small contractors: Buying tends to win out because work trucks take a beating, mileage adds up fast, and the ability to customize freely matters. The Section 179 deduction can also offset a large chunk of the purchase price in year one. Leasing can make sense if you want to preserve cash, keep trucks under warranty, and your annual mileage stays modest.
Want me to run some rough numbers for a specific truck or scenario?
Try it right now. Take something you'd normally ask AI — and add one line at the end telling it how to format the response.
Try it on something you're actually working on. Drop a comment on LinkedIn and tell me what you used it on.
Three tools worth having in your toolkit.