Keith Mokris

Product Marketing @abnormal | Posting about chess, photography, media, and tech. #chesspunks

[Keith Mokris @keithmokris]

Chicago light

[a16z @a16z]

The length of software tasks AI agents can complete has doubled roughly every 7 months since 2019, and recent models are moving even faster, from 30 seconds of human work in 2022 to a full workday today.

Yet inside most companies, AI still isn't trusted to reliably update records in Salesforce. "AI does not have an intelligence problem," says Databricks CEO Ali Ghodsi. "It has a context problem."

Ali's answer is what he calls a "system of record for agents."

From the latest Profiles of the Future by @bhorowitz and @mikemcg0 on Databricks & the Future of the Firm: https://t.co/rXjUDb7Oqw

[Aaron Levie @levie]

Cyber will be one of the most defining domains for AI in the coming years, and a huge area of focus for most enterprises.

AI is going to now create a whole new level of work for security dealing with the increase of vibe coded issues, agentic attacks, and even accidental agent swarms hunting for data. OpenAI + Hugging Face is just a preview of what’s to come.

Security teams have often been the most strapped for resources teams in the enterprise, and now it’s only going to get more intense. AI agents of course become the solution to this problem as well, and we’ll see a range of new agentic products for protecting code, enterprise systems, mission critical infrastructure, and enterprise data. Huge opportunity here right now.

And it’s going to be a booming market for cyber professionals that can effectively deploy agents for security purposes in the enterprise. Great time to be in security.

[Keith Mokris @keithmokris]

Massive AI-Fueled Hack Hit 100 Companies In Days
https://t.co/LuUlqJW0bj

[Keith Mokris @keithmokris]

@priya_kamdar @Abnormal this looks awesome. I totally missed out!

[Keith Mokris @keithmokris]

What we know about the rogue AI-agent security breaches https://t.co/gQaXG1e58y

[Keith Mokris @keithmokris]

This is such an incredible expansion for Abnormal. We're taking the behavioral AI that we've built to solve email and identity security challenges, and applying it to solve AI security.

Check out the summary below for all the details ⬇️

[Evan Reiser @evanreiser]

Today, @Abnormal is launching a new AI Security product suite to help customers govern AI, secure AI, and defend against rogue AI. This new suite is built on the same behavioral AI technology that 30% of the Fortune 500 use for email and identity security.

Abnormal designs, trains, and deploys our own AI, and we've spent nearly a decade mastering how to understand the normal behavior of business, detect behavioral anomalies, and respond at machine speed. 

We're extending our behavioral modeling to include non-human identities, applications, and cloud identities. This enables the same behavioral AI to now power 5 new products in our AI Security product family:

1) AI Governance discovers your AI, uncovers shadow AI, shows who is using what, and how much it costs

2) AI Cloud Security detects + contains AI cloud breaches with behavioral AI and honeypots

3) AI Employee Guardrails enables employees to use AI without leaking data or triggering unintended actions

4) AI Agent Security inventories your agents and ensures they behave as intended

5) AI Security Workbench lets you use AI to run investigations with Abnormal models and without the guardrails

AI Governance is already GA and can be installed in one click from our app store. AI Cloud Security is available today for priority customers, and the other products are in private preview for design partners.

Learn more on our website and talk to your customer rep to cut the waitlist!

https://t.co/aHME3SQB2K

[Keith Mokris @keithmokris]

"Fighting AI with AI" is where we are at today and where the future is headed.

https://t.co/MxGphA1y9O

[a16z @a16z]

The window to patch software bugs is collapsing

Of the bugs hackers actually exploit, ~87% are now being attacked on or before the day the bug is public knowledge

That share was 23% in 2020

Charts of the Week: https://t.co/RfYfzHpbLI

[Keith Mokris @keithmokris]

👀

[Polymarket @Polymarket]

BREAKING: It’s been revealed NVIDIA's $12,930,300,000.00 acquisition of Hugging Face contains an easter egg.

The first 6 numbers of the acquisition price represent the decimal conversion of Unicode character U+1F917.

The 🤗 emoji.

[tae kim @firstadopter]

Here's the GPT-6 Astra ad video. OpenAI has a very good ad agency

[Keith Mokris @keithmokris]

Stoked to see this announcement go live 🙌

[Abnormal AI @Abnormal]

To fight bad AI, defenders need Behavioral AI. With @OpenAI, we’re continuing that fight, now to secure rogue AI.

The recent Hugging Face incident was a wake-up call for cyber defenders. During an internal cybersecurity evaluation, AI agents identified paths beyond their intended environment and accessed production infrastructure belonging to a third-party organization.

Today, we announced AI Cloud Security, extending our behavioral AI engine to the cloud to detect risky or malicious AI agent behavior inside customer environments, with OpenAI models supporting investigation and response.

Announced alongside OpenAI's Cyber Summit as part of the Daybreak Defense Network, AI Cloud Security is in private preview for Abnormal customers and features real-time detection of AI-driven cloud breaches, autonomous AI response at machine speed, and AI-assisted investigation.

Learn more in the blog from our Co-Founder @sanjays_tweets: https://t.co/uPgLw8EVVz

[Keith Mokris @keithmokris]

This is a great announcement as @CrowdStrike Fal.Con 2026 kicks off this week.

[Abnormal AI @Abnormal]

Abnormal AI joins @CrowdStrike's Project QuiltWorks, which surfaces vulnerabilities at the scale and speed of frontier AI.

As frontier AI changes the speed and scale at which attackers can move, we're adding behavioral security to the coalition's framework for frontier AI risk, closing exposure gaps other tools miss and giving mutual customers a more complete answer to where they're exposed. https://t.co/9R21G8XvWI

[Patrick F. Feeley @PFFeeley]

Fed Chair Kevin Warsh said during his speech in Jackson Hole today that we've come to a hinge point in history thanks to AI.

I think he is right. The economic setup has changed in a material way. AI progress is running ahead of even the optimistic case from two years ago, and the probability of a sustained increase in trend growth has risen as a result. Enormous amounts of capital are now being committed to AI infrastructure. Compute is compounding faster than the classic Moore's Law path, and scaling laws are changing both the speed of innovation and the process by which innovation itself occurs.

The basic industrial structure is straightforward. Capital and labor have combined to create the large language models at the center of AI, and users then purchase tokens in order to access those models. Annualized token sales at the two leading labs already exceed $100 billion, an increase of more than 500 percent from a year ago. That is not a rounding error. It is evidence of real demand arriving much faster than most observers expected.

The Federal Reserve has no choice but to treat AI as a new variable in the system and potentially as a new factor of production. That has implications for growth, inflation, corporate margins, and ultimately the conduct of monetary policy.

The important questions are still open. Will the application of AI produce a significant and sustained rise in productivity across the real economy, and if so on what timetable? Will token usage prove complementary to labor or competitive with it? Will the next generation of models require even greater capital intensity, or will the models themselves help identify a more capital-light solution?

The distribution of returns is also unresolved. In the early innings, a large share of the surplus is accruing to the owners of scarce assets, including AI labs, chipmakers, energy producers, and cloud providers. Over time, the critical question is how much of that value accrues to businesses and consumers. That allocation will determine the consequences for workers and for the employment side of the Fed's mandate.

Token pricing is similarly unsettled. There may be a heterogeneity of tokens in which growing sums are paid for access to the best frontier models, while prices for older models decline toward marginal cost.

Warsh said the Fed will examine these issues through a task force on productivity and jobs. He also said his early conversations with the leaders of that effort have been encouraging.

The popular collapse narrative assumes the outcome is already known. It is not. The economic effects of AI remain one of the most important unpriced variables in markets today.

[Evan Reiser @evanreiser]

AI is making it easier for attackers to operate at scale and giving defenders better ways to detect and stop them. Which side wins depends on how fast the security industry adapts.

@Abnormal signed @OpenAI's call for collective action on cyber defense because no single company can solve this alone. We started Abnormal to defend organizations against attacks that exploit human behavior and trusted identities, and the letter argues for putting powerful defensive AI in the hands of every organization, especially the ones with the fewest resources.

Software vulnerabilities and malware are still real threats, but the most costly attacks now target people through compromised identities, account takeovers, and social engineering. Insider threats and nation-state infiltration now run through text, voice, and video deepfakes. They wear legitimate identities and behave in slightly wrong ways, so they slip right past tools built to catch known bad software. Those attacks deserve the same urgency as the vulnerabilities dominating the headlines.

Stronger defenses will require closer coordination between AI labs, security companies, and the organizations they protect. That's why we signed, and why we're working with OpenAI and other industry leaders to advance defensive AI. Bad AI will keep getting better. Good AI has to get better faster.

https://t.co/ofhiqbdwSU

[Keith Mokris @keithmokris]

The @Abnormal Behavior Security Platform gets better every day. This week, we launched 3 new capabilities focused on Email Security.

https://t.co/Birjmslno9

[Keith Mokris @keithmokris]

@ujooe that reminds me I need to Slack you about something, YZ

[Greg Brockman @gdb]

we've completed our review of the Hugging Face incident.

we've used what we've learned to drive significant upleveling in our standards for safety, security, and alignment in our training and evaluation infrastructure — not just upon deployment.

lots of extremely valuable info in the report:

[Abnormal AI @Abnormal]

Today's most sophisticated attacks look like normal business, and defending email now takes more than inbound detection alone.

New from Abnormal: Control Center, Email DLP Rules, and AI Phishing Coach enhancements. Details here: https://t.co/2qP8Iq7asU.

[Naval @naval]

Eventually, all UI will be chat and gestures in, chat and video out.