StarZero All shows Try the podcast clipper →

Jun 8, 2026 · 31m · allin

Palo Alto Networks CEO: "AI Found 5 Years of Bugs in 6 Weeks"

Nikesh Arora · 18m spoken Chamath Palihapitiya · 3m spoken Jason Calacanis · 2m spoken David Friedberg · 1m spoken David Sacks · 48s spoken
0:00 / 0:00
▶ Watch on YouTube →

gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions

At the All-In Liquidity Summit 2026, Palo Alto Networks Chairman and CEO Nikesh Arora joins the All-In Podcast hosts to discuss enterprise AI transformation, cybersecurity challenges, and the changing landscape of software business models.

How this conversation actually went

Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. The hosts hold 31.6% of the talking time here. How this is scored →

The hosts as informed peer 4.3 Guest teaching 5.8 Guest disagreement 3.5 The hosts pushing back 2.8
05100:0010:0020:0030:000:30–2:33 · The hosts as informed peer 4/10 Nikesh Arora's Tenure and Palo Alto Networks Growth Chamath sets up the interview by citing Palo Alto Networks market cap growth from $17B to $238B and framing their path to $1T. Nikesh warmly accepts the praise and introduces how AI democratizes operational intelligence across enterprise teams.2:33–6:50 · The hosts as informed peer 4/10 AI Vulnerability Discovery with Mythos Nikesh reveals that Mythos found code vulnerabilities in 6 weeks that normally take 5-7 years, surprising the hosts. Sacks frames the cyber defense vs attack race while Chamath briefly probes whether the capabilities are overhyped, which Nikesh quickly corrects.6:50–11:32 · The hosts as informed peer 5/10 Death of Analytical SaaS and UI Elimination Nikesh delivers a sweeping reframe of the software industry, declaring analytical SaaS dead and UI to be the worst design legacy of tech. Jason validates this with his own company's 90% SaaS bill reduction, while Chamath prompts Nikesh on vendor license disputes.11:32–14:05 · The hosts as informed peer 3/10 National Security Risks and Cyber Attack Realities When Chamath raises high-level national security risks from foreign state actors, Nikesh reframes the premise by pointing out that 89% of breaches stem from basic credential theft rather than sophisticated model exploits, emphasizing small business vulnerability over grid attacks.14:05–18:29 · The hosts as informed peer 5/10 Utility AI Models and Application Layer Profit Pools Nikesh argues that AI models are becoming a commoditized utility layer and that profit pools reside in applications. Jason pushes back using the Microsoft Office bundling analogy, but Nikesh firmly maintains that enterprises will buy specialized application planes rather than building on raw models.18:29–25:30 · The hosts as informed peer 5/10 High False Positive Rates and Enterprise Reliability Nikesh challenges the hosts for ignoring model false-positive rates, revealing Mythos had a 30% false-positive rate which makes raw models unusable for enterprise defense. He then playfully calls out Friedberg for previously distinguishing between founder and non-founder CEOs, forcing Friedberg into a diplomatic retreat.0:30–2:33 · Guest teaching 3/10 Nikesh Arora's Tenure and Palo Alto Networks Growth Chamath sets up the interview by citing Palo Alto Networks market cap growth from $17B to $238B and framing their path to $1T. Nikesh warmly accepts the praise and introduces how AI democratizes operational intelligence across enterprise teams.2:33–6:50 · Guest teaching 6/10 AI Vulnerability Discovery with Mythos Nikesh reveals that Mythos found code vulnerabilities in 6 weeks that normally take 5-7 years, surprising the hosts. Sacks frames the cyber defense vs attack race while Chamath briefly probes whether the capabilities are overhyped, which Nikesh quickly corrects.6:50–11:32 · Guest teaching 7/10 Death of Analytical SaaS and UI Elimination Nikesh delivers a sweeping reframe of the software industry, declaring analytical SaaS dead and UI to be the worst design legacy of tech. Jason validates this with his own company's 90% SaaS bill reduction, while Chamath prompts Nikesh on vendor license disputes.11:32–14:05 · Guest teaching 6/10 National Security Risks and Cyber Attack Realities When Chamath raises high-level national security risks from foreign state actors, Nikesh reframes the premise by pointing out that 89% of breaches stem from basic credential theft rather than sophisticated model exploits, emphasizing small business vulnerability over grid attacks.14:05–18:29 · Guest teaching 6/10 Utility AI Models and Application Layer Profit Pools Nikesh argues that AI models are becoming a commoditized utility layer and that profit pools reside in applications. Jason pushes back using the Microsoft Office bundling analogy, but Nikesh firmly maintains that enterprises will buy specialized application planes rather than building on raw models.18:29–25:30 · Guest teaching 7/10 High False Positive Rates and Enterprise Reliability Nikesh challenges the hosts for ignoring model false-positive rates, revealing Mythos had a 30% false-positive rate which makes raw models unusable for enterprise defense. He then playfully calls out Friedberg for previously distinguishing between founder and non-founder CEOs, forcing Friedberg into a diplomatic retreat.0:30–2:33 · Guest disagreement 1/10 Nikesh Arora's Tenure and Palo Alto Networks Growth Chamath sets up the interview by citing Palo Alto Networks market cap growth from $17B to $238B and framing their path to $1T. Nikesh warmly accepts the praise and introduces how AI democratizes operational intelligence across enterprise teams.2:33–6:50 · Guest disagreement 2/10 AI Vulnerability Discovery with Mythos Nikesh reveals that Mythos found code vulnerabilities in 6 weeks that normally take 5-7 years, surprising the hosts. Sacks frames the cyber defense vs attack race while Chamath briefly probes whether the capabilities are overhyped, which Nikesh quickly corrects.6:50–11:32 · Guest disagreement 3/10 Death of Analytical SaaS and UI Elimination Nikesh delivers a sweeping reframe of the software industry, declaring analytical SaaS dead and UI to be the worst design legacy of tech. Jason validates this with his own company's 90% SaaS bill reduction, while Chamath prompts Nikesh on vendor license disputes.11:32–14:05 · Guest disagreement 5/10 National Security Risks and Cyber Attack Realities When Chamath raises high-level national security risks from foreign state actors, Nikesh reframes the premise by pointing out that 89% of breaches stem from basic credential theft rather than sophisticated model exploits, emphasizing small business vulnerability over grid attacks.14:05–18:29 · Guest disagreement 5/10 Utility AI Models and Application Layer Profit Pools Nikesh argues that AI models are becoming a commoditized utility layer and that profit pools reside in applications. Jason pushes back using the Microsoft Office bundling analogy, but Nikesh firmly maintains that enterprises will buy specialized application planes rather than building on raw models.18:29–25:30 · Guest disagreement 5/10 High False Positive Rates and Enterprise Reliability Nikesh challenges the hosts for ignoring model false-positive rates, revealing Mythos had a 30% false-positive rate which makes raw models unusable for enterprise defense. He then playfully calls out Friedberg for previously distinguishing between founder and non-founder CEOs, forcing Friedberg into a diplomatic retreat.0:30–2:33 · The hosts pushing back 1/10 Nikesh Arora's Tenure and Palo Alto Networks Growth Chamath sets up the interview by citing Palo Alto Networks market cap growth from $17B to $238B and framing their path to $1T. Nikesh warmly accepts the praise and introduces how AI democratizes operational intelligence across enterprise teams.2:33–6:50 · The hosts pushing back 3/10 AI Vulnerability Discovery with Mythos Nikesh reveals that Mythos found code vulnerabilities in 6 weeks that normally take 5-7 years, surprising the hosts. Sacks frames the cyber defense vs attack race while Chamath briefly probes whether the capabilities are overhyped, which Nikesh quickly corrects.6:50–11:32 · The hosts pushing back 2/10 Death of Analytical SaaS and UI Elimination Nikesh delivers a sweeping reframe of the software industry, declaring analytical SaaS dead and UI to be the worst design legacy of tech. Jason validates this with his own company's 90% SaaS bill reduction, while Chamath prompts Nikesh on vendor license disputes.11:32–14:05 · The hosts pushing back 3/10 National Security Risks and Cyber Attack Realities When Chamath raises high-level national security risks from foreign state actors, Nikesh reframes the premise by pointing out that 89% of breaches stem from basic credential theft rather than sophisticated model exploits, emphasizing small business vulnerability over grid attacks.14:05–18:29 · The hosts pushing back 4/10 Utility AI Models and Application Layer Profit Pools Nikesh argues that AI models are becoming a commoditized utility layer and that profit pools reside in applications. Jason pushes back using the Microsoft Office bundling analogy, but Nikesh firmly maintains that enterprises will buy specialized application planes rather than building on raw models.18:29–25:30 · The hosts pushing back 4/10 High False Positive Rates and Enterprise Reliability Nikesh challenges the hosts for ignoring model false-positive rates, revealing Mythos had a 30% false-positive rate which makes raw models unusable for enterprise defense. He then playfully calls out Friedberg for previously distinguishing between founder and non-founder CEOs, forcing Friedberg into a diplomatic retreat.

speaking balance: gold is the hosts, purple is the guest (3 minute bins)

0:00 · the hosts 28.9% · guest 71.1%0:00 · the hosts 28.9% · guest 71.1%3:00 · the hosts 32.6% · guest 67.4%3:00 · the hosts 32.6% · guest 67.4%6:00 · the hosts 27.7% · guest 72.3%6:00 · the hosts 27.7% · guest 72.3%9:00 · the hosts 34.8% · guest 65.2%9:00 · the hosts 34.8% · guest 65.2%12:00 · the hosts 19.7% · guest 80.3%12:00 · the hosts 19.7% · guest 80.3%15:00 · the hosts 31.8% · guest 68.2%15:00 · the hosts 31.8% · guest 68.2%18:00 · the hosts 37% · guest 63%18:00 · the hosts 37% · guest 63%21:00 · the hosts 50.2% · guest 49.8%21:00 · the hosts 50.2% · guest 49.8%24:00 · the hosts 22.3% · guest 77.7%24:00 · the hosts 22.3% · guest 77.7%27:00 · the hosts 37.6% · guest 62.4%27:00 · the hosts 37.6% · guest 62.4%30:00 · the hosts 18.3% · guest 81.7%30:00 · the hosts 18.3% · guest 81.7%
Sharpest disagreement ▶ 18:52 Nikesh confronts hosts over ignoring false positive rates

Nikesh explicitly calls out the hosts' tech-optimist blind spot, telling them directly that they fail to talk about false positive rates and explaining why a 30% error rate renders raw models useless for enterprise security.

 Hardest push from the hosts ▶ 15:25 Jason challenges Nikesh using the Microsoft Office expansion analogy

Jason refuses Nikesh's framing that model providers will stay out of the application layer, pressing him on whether Anthropic and OpenAI will follow Microsoft's playbook of releasing applications directly on top of their operating platform.

 Biggest teaching moment ▶ 18:52 Nikesh educates hosts on false positive rates in enterprise deployment

Nikesh resets the conversation around model benchmarks by revealing Mythos's 30% false positive rate, demonstrating that consumer model enthusiasm fails in enterprise defense where tolerance for false alarms is near zero.

 The host holds their own ▶ 7:46 Jason demonstrates real-world SaaS reduction via API integration

Jason proves his domain expertise by explaining how his team eliminated 17 of 20 SaaS seats by connecting their underlying data store directly to Slack and Claude, resulting in an immediate 90% cost savings.
the scores for every segment, with the reasoning behind each
ChapterTopicThe hosts as informed peerGuest teachingGuest disagreementThe hosts pushing backWhy
Nikesh Arora's Tenure and Palo Alto Networks Growth 4311 Chamath sets up the interview by citing Palo Alto Networks market cap growth from $17B to $238B and framing their path to $1T. Nikesh warmly accepts the praise and introduces how AI democratizes operational intelligence across enterprise teams.
AI Vulnerability Discovery with Mythos 4623 Nikesh reveals that Mythos found code vulnerabilities in 6 weeks that normally take 5-7 years, surprising the hosts. Sacks frames the cyber defense vs attack race while Chamath briefly probes whether the capabilities are overhyped, which Nikesh quickly corrects.
Death of Analytical SaaS and UI Elimination 5732 Nikesh delivers a sweeping reframe of the software industry, declaring analytical SaaS dead and UI to be the worst design legacy of tech. Jason validates this with his own company's 90% SaaS bill reduction, while Chamath prompts Nikesh on vendor license disputes.
National Security Risks and Cyber Attack Realities 3653 When Chamath raises high-level national security risks from foreign state actors, Nikesh reframes the premise by pointing out that 89% of breaches stem from basic credential theft rather than sophisticated model exploits, emphasizing small business vulnerability over grid attacks.
Utility AI Models and Application Layer Profit Pools 5654 Nikesh argues that AI models are becoming a commoditized utility layer and that profit pools reside in applications. Jason pushes back using the Microsoft Office bundling analogy, but Nikesh firmly maintains that enterprises will buy specialized application planes rather than building on raw models.
High False Positive Rates and Enterprise Reliability 5754 Nikesh challenges the hosts for ignoring model false-positive rates, revealing Mythos had a 30% false-positive rate which makes raw models unusable for enterprise defense. He then playfully calls out Friedberg for previously distinguishing between founder and non-founder CEOs, forcing Friedberg into a diplomatic retreat.

Statements from this episode (17)

Prediction Open · timeframe Jun 2031
Palihapitiya: Palo Alto Networks on track for $1 trillion market cap
“So you're on your way to a trillion dollars.”
Chamath Palihapitiya Jun 8, 2026 ▶ 0:50
Insight
Arora: AI enables 90% output consistency across Palo Alto's marketing team
“There's this notion, which we talked about briefly last time, that AI is really democratizing intelligence. What that means is I have 250 people in marketing that produce varied forms of output. Now I can get 90% of the output to be consistent across those 250…”
Nikesh Arora Jun 8, 2026 ▶ 1:42
Assertion Not checkable as stated
Arora: AI found five years of software vulnerabilities in six weeks
“In six weeks we found vulnerabilities which would have normally taken us five to seven years to find.”
Nikesh Arora Jun 8, 2026 ▶ 2:55
Prediction Held up
Arora: Vulnerability-finding AI will be public within three months
“I think we're three months away, if not already there, from this being available in the wild.”
Nikesh Arora Jun 8, 2026 ▶ 4:30
Insight
Arora: Enterprises must collect 10x more cybersecurity data against AI
“We need to collect 10 times the data in the enterprise from a cyber perspective to be able to understand how to defend ourselves against the AI attackers.”
Nikesh Arora Jun 8, 2026 ▶ 6:40
Prediction Not checkable as stated
Nikesh Arora: Analytical SaaS companies are over
“If you're an analytical SaaS company, it's over.”
Nikesh Arora Jun 8, 2026 ▶ 7:08
Disclosure
Calacanis: Cut SaaS spending 90% using Claude and Slack integrations
“We had an instance with a SAS product with 20 seats. Nobody was logging in and using it, but the data was there. So we created like three accounts, got rid of 17, connected it to Slack, connected it to Claude. And now everybody can interface it through a natur…”
Jason Calacanis Jun 8, 2026 ▶ 7:48
Prediction Open · timeframe Jun 2029
Nikesh Arora: Enterprise data storage needs will grow 10x in three years
“We are going to need 10 times the data stored enterprise than we have today.”
Nikesh Arora Jun 8, 2026 ▶ 9:10
Prediction Not checkable as stated
Nikesh Arora: Systems of record software will be reinvented within five years
“So I think the whole system of work System of record gets reinvented the next five years.”
Nikesh Arora Jun 8, 2026 ▶ 11:25
Assertion Contradicted
Arora: 89% of cyber attacks occur because credentials get stolen
“89% of the attacks happen because credentials get stolen.”
Nikesh Arora Jun 8, 2026 ▶ 12:11
Opinion
Arora: AI threats increase the terminal value of cybersecurity
“I just think it increased the terminal value of the industry, right?”
Nikesh Arora Jun 8, 2026 ▶ 13:40
Prediction Not checkable as stated
Arora: AI base models will become a commoditized utility layer
“I still believe models are going to become a utility layer. You'll be able to buy intelligence on the fly”
Nikesh Arora Jun 8, 2026 ▶ 14:13
Assertion Not checkable as stated
Arora: Frontier AI model weights fit on a single USB stick
“The entire model weights of their newest model fits on a USB stick.”
Nikesh Arora Jun 8, 2026 ▶ 18:05
Assertion Not checkable as stated
Arora: Claude Mythos had a 30% false positive rate in testing
“The false positive rate on Mythos was 30%.”
Nikesh Arora Jun 8, 2026 ▶ 19:06
Prediction Open · timeframe Jun 2031
Arora: Google will be the first $10 trillion company
“I think Google's underrated. I think it's going to be the first 10 trillion dollar company in our lifetime.”
Nikesh Arora Jun 8, 2026 ▶ 21:39
Prediction Not checkable as stated
Arora predicts 10 years for US to fulfill hardware supply chain needs
“In 10 years.”
Nikesh Arora Jun 8, 2026 ▶ 27:39
Disclosure
Arora: AI is driving record tech headcount at Palo Alto Networks
“I think we're gonna have more people at Palo Alto on the technology side than we've ever had before, because I think AI is causing everything to ask for a transformation.”
Nikesh Arora Jun 8, 2026 ▶ 30:51
Made with StarZero

Turn any episode into a week of clips.

This entire site, over 460 episodes transcribed, diarized, checked and made playable, runs on the StarZero media pipeline. Drop in your own episode and the podcast clipper finds the moments worth sharing, cuts them, captions them, and reframes them for every feed.

Try the podcast clipper Learn more