It's "AI-enabled tradecraft" as in let's take a guess at Vercel leadership's pressure to install and test AI across the company, regardless of vendor risk? Speed speed speed.

This is an extremely vanilla exploit that every company operating without a strictly enforceable AI install allowlist is exposed to - how many AI tools like Context are installed across your scope of local and SaaS AI? Odds are, quite a bit, or ask your IT guy/gal for estimates.

These tools have access to... everything! And with a security vendor and RBAC mechanism space that'll exist in about... 18-24 months.

Vercel is the canary. It's going to get interesting here, no way in heck that Context is the only target. This is a well established, well-concerned/well-ignored threat vector, when one breaks open the other start too.

Implies a very challenging 6 months ahead if these exploits are kicking off, as everyone is auditing their AI installs now (or should be), and TAs will fire off with the access they have before it is cut.

Source - am a head of sec in tech

?

So, what can derail AI out of left field? Maybe building DCs for it in Arizona and EMEA can, for one.... choosing very "water-rich" locations there for water-cooled systems.

So, how could this land longterm, assuming AI works sort of good, sort of bad against the use cases? The real questions here for industry people though should be this:

1) How does this play out, over the 5-10 yrs we have to see it occur of trying it/redoing it/trying a new version/going back to the old version, all the while it's occurring over my career, all the while when I have bills to pay and relationships to maintain.

Ans: I think that's a hell of a lot of financial and employment stress induced on us by people who don't understand the tech they're rolling out, the state change that's occurring, and don't need to deal with the consequences. All the while, I go mid career, to late career, dealing with what AI can actually do in the background.

2) What is actually going to work wrt being relevant to my job?

Ans: I think what actually works is the vuln research aspect of AI, feedback loops rapidly, rapidly speeding up on that.

And, what is the most stressful, obnoxious, high burnout part of the job - sec arch and vuln remediation, or IR and vuln response. Both about to go on overddrive, and already are if you're minding bug bounties and IR these days.

3) Has this happened to other industries, how did it go?

Ans: trading, trading, trading, trading. Check it out.

And if you're planning it, plan it soon b/c vendors like Dropzone are carving out the entry sec eng ops/ir jobs in-house or at the MSPs, and Trail of Bits skills foss on GH are carving out the 2-3x extra $3-400k TC line sec eng roles .

Ya I get the need but you miss the point - no, you can't pay me anymore to wade into that and own risk, beyond a consulting context with low skin in the game.

There is a wave of senior leads thinking like this, because the knife's edge of "enough risk to game it for pay" finally tilted too far, and the career has changed.

In terms of going home after work and not yelling at my kids and spouse due to work stress due to the 10th 0day in a week on my corporate VPN/my retail-facing app/my..., there's a real QoL issue to consider. Many outside of security consistently misunderstands the mental health/career satisfaction/pay triad.

More or less, I am the attractive resume, and: the game has changed folks.

For what it is worth, I am taking my ball and going home in about 12 months. I've saved enough, locked in a perma-middle class lifestyle in a great nondescript city, and swapping over to offensive consulting and a AI-free, non-tech trade that won't take too long to get into - think a PA, nurse, plumber, etc.

I'm not quite old enough and with the end of responsibilities as to FIRE, but I can read the writing on the wall enough to understand an AI-proof FI needs to be locked in before everyone else realizes the same. Many others in sec are feeling this.

I think tech will find security pros willing to throw themselves into the fray for pay and optimism. There are others like me who are extracting their final nuts. There are others who have golden-handcuffed themselves into this ride with their mortgages and private school tuitions. And I'm sure some others will stick it out. There will also be an AI-enabled version of sec eng soon enough.

But if private sector doesn't wake up to AI integrations - internal doc rollouts hoovering up PII that wasn't supposed to be stored there, externally-facing customer support portals social engineered and pivoted into, PRs via Slack comment via marketing hires who are ATO'd - this is going to be a 1990's-style BBQ where 0days on critical systems are dropped at happy hours at conferences nightly.

And: your security teams are going to be burned out, banking up, and quitting. The risk acceptances, the double-speak, the slow-rolling, the half-baked risk thinking for engineering and product leads, the corners cut, the public endpoints opened up just this one time - that's going to be enough rope, and already is enough, to hang yourself in this offensive context that's building now.

It is deeply humorous that SWE and engineering leadership has worked itself into this position via its AI push to unemploy itself while thinking it's the 1x white collar job exempt from automation threats.

All it'll take is another recession like '08, and the leaves get shaken off the trees finally. Thankfully there is only one (wait, there are two probably), thankfully there are only two-to-three (wait, there are like 10) systemic market threats right now.