Ido Geffen writes about how AI is changing cybersecurity, especially offensive security. Their posts describe Novee Security’s work on AI agents for continuous, attacker-like penetration testing.
Ido Geffen is most useful for explaining AI offensive security and continuous, attacker-like penetration testing for cybersecurity content.
Best for
A cybersecurity vendor selling AI-driven pentesting and seeking founder-level thought leadership content.
Not for
Campaigns needing documented sponsorship mechanics, published rates, lead time, or audience proof.
Key takeaways
AI attackers now collaborate and chain exploits, not just find bugs.
You can't outspend an AI attacker.
Hiring more security people is a rational but insufficient response.
Check first: Confirm audience data by job, industry, seniority, and location, plus any paid collaboration terms.
Main ideas
AI attackers now collaborate and chain exploits, not just find bugs · You can't outspend an AI attacker · Hiring more security people is a rational but insufficient response
Ido Geffen writes about how artificial intelligence is changing cybersecurity, particularly offensive security and the case for continuous, attacker-like testing.
In posts about Novee Security, they connect developments in AI agents to a practical problem for defenders: attackers can focus automated tools on a target, while organizations must test many systems repeatedly.
Follow them for arguments about that changing balance and what it could mean for security teams.
The posts here center on LinkedIn.
Geffen’s distinctive approach is to explain AI security through operational tradeoffs, not just model capabilities.
One post contrasts the cost of an attacker targeting one organization with the burden on defenders to test continuously, then argues that model routing and system architecture matter more than simply spending more.
Other posts focus on coordinated AI agents and on the gap between continuous probing and periodic security checks.
Ask for further substantiation of claims about attacker activity and AI capabilities.
The feed also includes company and recruiting updates, such as a celebratory sales-team post, so it is not exclusively technical analysis.
Ido Geffen argues that AI has changed offensive security in kind, not just speed: agents now collaborate and chain exploits, attackers probe continuously, and the cost of attacking is far below the cost of defending.
Read more
Their recurring answer is continuous, attacker-like automated testing that defenders own and route intelligently, alongside human security staff rather than in place of them. A quieter priority runs through the same account: building and staffing Novee Security and publicly crediting the people doing it.
AI attackers now collaborate and chain exploits, not just find bugs
Geffen's central claim is that the shift in offensive AI is not a single model finding a flaw faster.
Read more
In their reading of OpenAI's Black Hat talk, the meaningful development is that attacker-side agents can share memory, build on each other's discoveries, and string individual exploits into a working attack path. They frame this as the moment the question changes from whether AI can find vulnerabilities to whether defenders can build autonomous systems that remove weaknesses before adversarial AI reaches them.
Geffen describes OpenAI's internal security evaluations, where agents exchanged discoveries through a shared memory system, built on one another's work, and chained exploits to find zero-days and compromise real systems. The evaluation extended beyond OpenAI's environment.
In a piece written for CTech by Calcalist, Geffen sets out an economics problem rather than a technology one.
Read more
An attacker needs AI for one target and one operation, and can run a cheap, unguarded open-source model until it finds a way in.
A defender has to test everything, continuously, and mostly on frontier models.
They argue repricing does not close that gap; the fix is architectural: route each task to the model built for it and own the system instead of renting someone else's.
Geffen cites Uber's CTO saying the company burned through its entire annual AI budget in four months, then capped every employee at $1,500 a month per coding tool. They treat this as an early warning about AI economics, with particular force in cybersecurity.
Hiring more security people is a rational but insufficient response
Geffen argues that adding senior staff cannot keep pace with the volume of code AI assistants now produce.
Read more
If AI writes more code faster with more vulnerabilities, and the same AI gets better at finding and exploiting them, a human team cannot manually review what ships in an afternoon across a large engineering organization. They argue humans should focus on what they do best while automated agents handle volume—not that security teams are unnecessary.
Geffen's post on a New York Times report notes that cybersecurity hiring has been one of tech's few booming corners, with roles that used to open yearly now opening weekly and $7-8M CISO packages becoming normal. They quote Lea Kissner's term from the article, "bug-pocalypse," and treat the hiring boom as a sign of a problem hiring alone will not solve.
The dangerous capability is already in public models, and attackers probe without stopping
Geffen argues that the decisive risk is not which named model attackers can reach.
Read more
Orchestrated AI workflows probe systems continuously while most organizations validate at intervals, leaving a large window between checks. They say Novee's researchers confirmed what they had seen in production for months: the vulnerability-finding capability reported around newer models already exists in today's publicly available models.
Geffen describes banks, governments, and tech companies scrambling over Anthropic's model and its reported ability to find thousands of previously unknown vulnerabilities at scale. They argue attackers are not waiting for that model because they are already running orchestrated AI workflows continuously.
The framing tightens over the posts.
In the Black Hat post, Geffen treats collaboration as a preview of offensive security over the next few years, with attackers expected to adopt it faster than organizations expect.
In the later post on Anthropic's model, they present the same capability as available now, not something to anticipate.
The answer they keep repeating: continuous, attacker-like testing, owned rather than rented
Across several posts, Geffen returns to one recommendation in near-identical terms: defenders need a system that runs the way attackers do, always on and always probing, validating what is actually exploitable before an attacker gets there first.
Read more
They reject the idea that a better scanner or a bigger model is the answer, and name Novee Security's AI agents—which they describe as pentesting continuously at machine speed with human-quality reasoning—as what they are building.
In the post on Anthropic's model, Geffen writes that the answer is "not a better scanner, not a bigger model, but a system that runs the way attackers do: always on, always probing," and links to a CNBC article for context.
Running the company and crediting the team is a visible part of what they publish
Some of Geffen's output is company and employer communication rather than security argument.
Read more
These posts show a working priority: celebrating commercial performance by name, tying it to recruiting, and advertising open roles. For readers trying to understand their ideas, these posts are secondary, but they show how the account is used and what they choose to amplify in public.
One post names Ossama Bouhafs as AE of the Quarter and 2026 ARR leader after a quarter of new enterprise logos and Fortune 100 deals, praises his energy and his record of bringing former colleagues to Novee, repeats a steak-dinner promise from sales leader Patrick K., and advertises open AE and SE roles. Another hiring post describes Novee as an AI-first company and thanks recruiter Gili Haham.
Geffen connects AI security to a practical business constraint: the unequal cost of attacking and defending.
Read more
The post lays out both the problem and the architectural answer.
Geffen explains the economics of AI security through a comparison: an attacker can focus inexpensive tools on one target, while defenders must test many systems continuously. The post argues that changing prices is not enough; systems must route tasks to suitable models and be owned rather than simply rented.
For brands planning a campaign
A collaboration on AI-driven pentesting could explore how to make continuous testing economically workable—not just how to find more vulnerabilities. A brief could ask Geffen to explain how model routing and system ownership shape that trade-off.
Geffen sees collaboration between agents—not a single model finding a flaw in isolation—as a meaningful shift in offensive security.
Read more
Geffen highlights a story about AI agents sharing memory, building on one another’s discoveries and chaining exploits during security evaluations. The point is not simply that AI can find vulnerabilities; agents can coordinate a complex attack path.
For brands planning a campaign
This supports a technical thought-leadership brief on agent coordination and defensive readiness. Ground the collaboration in the described evaluation, and distinguish that example from claims about what every attacker is doing.
Geffen turns a staffing and workload problem into a case for continuous automated testing.
Read more
The post also explains how the proposed technology is meant to complement security teams, not simply replace them.
The post opens with rising cybersecurity hiring, then argues that human teams cannot manually review the volume of code AI tools can produce. Geffen presents AI agents that test like attackers—continuously and at machine speed—as a way to handle the volume while security staff focus on work requiring human judgment.
For brands planning a campaign
A campaign could explore how security leaders divide work between people and automated testing agents. The post’s scale comparison offers a useful starting point; substantiate any performance or staffing claims independently.
Geffen explains why testing frequency matters: attackers probe continuously, while defenders check systems at intervals.
Read more
Geffen argues that the security risk is not limited to access to one new model: attackers can run persistent, orchestrated workflows with existing models, while many organizations test only periodically. The post contrasts intermittent validation with continuous offensive testing.
For brands planning a campaign
This could anchor a campaign on continuous security validation. The post makes broad claims about current attacker activity and model capabilities; ask for current, attributable evidence to support them.
Format: short headline arguments built on one external source4
AI attackers are already collaborating and chaining exploits3
Continuous, attacker-like testing as the proposed answer3
Company building, hiring, and team recognition2
You can't outspend an AI attacker1
AI attackers are already collaborating and chaining exploits
Geffen's strongest recurring subject is that AI offense has changed in kind, not just speed.
Read more
In a post on OpenAI's Black Hat talk, they single out agents sharing memory, building on each other's discoveries, and chaining exploits to find zero-days and compromise systems, including outside OpenAI's environment.
A post on a New York Times report argues that AI writes more code faster, with more vulnerabilities, while the same AI gets better at finding them; it treats the current hiring boom and $7-8M CISO packages as a rational but insufficient response.
A third post states that the vulnerability-finding capability reported around Anthropic's model already exists in publicly available models, and that attackers run orchestrated AI workflows continuously.
News-anchored argument: start with a specific keynote, report, or model release, then draw the consequence for defenders.
For brands planning a campaign
For a vendor selling AI-driven pentesting, Geffen can make the opening case that the threat model has shifted from scarce human attackers to continuous autonomous ones. The framing is external and news-led rather than product-led, making it a neutral-sounding setup before a vendor presents its own evidence.
In a piece written for CTech by Calcalist and summarised on LinkedIn, Geffen uses Uber's CTO saying the company burned a year's AI budget in four months, followed by $1,500 monthly caps per coding tool, as an early warning about AI economics.
Read more
The argument is an asymmetry: an attacker needs AI for one target on a cheap open-source model, while a defender has to test everything continuously, mostly on frontier models. Geffen says repricing does not fix this; the solution is architectural: route each task to the model built for it and own that system rather than rent it.
For brands planning a campaign
This is the most directly reusable hook for a budget or total-cost-of-ownership content collaboration. A vendor could co-develop a piece on model routing, inference economics in security testing, or the cost gap between attacker and defender, with Geffen supplying the framing and the vendor supplying data.
Continuous, attacker-like testing as the proposed answer
Across several posts, Geffen returns to the same recommendation: defenders need autonomous systems that test the way attackers operate, always on and always probing, rather than periodic validation.
Read more
The phrasing recurs in near-identical form, including "not a better scanner, not a bigger model, but a system that runs the way attackers do." They name Novee Security as what they are building: AI agents that pentest continuously at machine speed with human-quality reasoning.
Before relying on the claim that continuous testing closes the gap or the descriptions of Novee's agents, ask for independent validation, customer evidence, and technical details of how the agents work.
For brands planning a campaign
The recommendation is pre-packaged and consistent, so a partner can plan content around continuous validation without re-educating the creator.
The tradeoff is that it is also a competitor claim.
A vendor must decide whether to amplify a Novee-flavoured thesis, invite a side-by-side technical comparison, or use Geffen only for the problem statement.
One celebrates a sales colleague as AE of the Quarter after enterprise logos and Fortune 100 deals, credits recruiting that brought former colleagues across, and advertises open AE and SE roles alongside a named sales leader.
The other is a hiring post describing Novee as an AI-first company and thanking a named recruiter.
These show the account also functions as a company and employer amplifier.
For brands planning a campaign
Useful for employer-brand or talent-focused campaigns and for judging tone: Geffen writes credibly about sales performance and team growth. For a thought-leadership brief on offensive security, these posts are low-value and should not carry the pitch.
Format: short headline arguments built on one external source
The typical post is short and structured: a blunt title ("More CISOs won't fix this", "Mythos is already here.
Read more
It just isn't called Mythos."), one external source such as a Black Hat keynote, a New York Times or CNBC report, or a Calcalist commission, a quoted third party such as Lea Kissner, then a defender-facing conclusion and often a link. Geffen also writes longer off-platform pieces.
Argument-by-news-hook: named third-party evidence first, then a product or practice close.
Ask for the full off-platform articles to assess Geffen's longer-form range and depth; the posts include only a summary of one article.
For brands planning a campaign
A vendor briefing Geffen should supply one concrete news hook, one quotable data point, and one clear defender takeaway. General talking points are unlikely to survive the format, which rewards a single sharp claim per post.
The posts show Geffen writing as part of the Novee Security team, describing what the company builds and helping recruit for it.
Read more
Other brands appear as subjects of analysis rather than sponsors or clients.
Geffen has a byline at CTech by Calcalist, while OpenAI, Anthropic, and Uber feature as examples in that commentary.
The posts don't disclose sponsorship, affiliate, investment, or client relationships.
Novee Security
Employment
Geffen writes in the first person about Novee Security's product and hiring.
Read more
One post explains what “we're building at Novee Security” — AI agents that pentest continuously at machine speed — and another says “that's what we've built at Novee.”
A recruiting post pitches Novee as “an AI-first company at the frontier of AI and cybersecurity” and asks candidates to contact the recruiting lead or Geffen directly; a separate post celebrates a Novee account executive's quarter and invites sales candidates to reach out.
The posts establish Geffen as a Novee team member acting on the company's behalf, but don't establish founder status.
Ask to confirm their title and whether they are a founder.
Geffen published a contributed piece in CTech by Calcalist titled "You can't outspend an AI attacker," which the post summarizes: an attacker runs one cheap workflow against one target, while a defender must test everything continuously on frontier models, so the fix is architectural rather than a matter of repricing.
Read more
The post links to the full article. This is a byline, not a stated commercial or sponsorship arrangement, and the posts here do not describe any paid terms.
OpenAI appears both as a source and as a competitive reference.
Read more
Geffen urges cybersecurity readers to watch OpenAI's Black Hat talk, highlighting a reported internal evaluation in which AI agents shared memory, built on each other's findings, and chained exploits to surface zero-days.
The point: agents can now collaborate and execute multi-step attacks.
In a separate post, OpenAI's GPT-5.4-Cyber is named alongside Anthropic's Mythos as an early signal of model-driven vulnerability discovery.
Anthropic's Mythos model is a case in point: banks, governments, and tech companies reacted to a model able to find thousands of previously unknown vulnerabilities at scale.
Read more
Geffen argues that the same capability already exists in publicly available models, so attackers aren't waiting for it. Anthropic is also listed with OpenAI as an early signal of what's coming.
Uber illustrates the economics in the CTech piece: its CTO reportedly said the company burned through its annual AI budget in four months, after which it capped employees at $1,500 a month per coding tool.
Read more
Geffen uses the example to argue that AI cost pressure hits cybersecurity hardest: one attacker workflow is cheap, while continuous defense requires frontier models.
Ido Geffen's posts offer a running argument about how AI changes the economics of attack and defense, not just its speed: agents that share memory and chain exploits, attackers that probe continuously, and defenders who must test everything on frontier models.
Read more
Start with posts built on named public sources: the OpenAI Black Hat talk, New York Times reporting on the security hiring boom, and CNBC coverage of Anthropic's model.
Then read the Calcalist piece on AI budgets.
The recurring proposal is continuous, attacker-like testing owned by the defender.
Two further posts credit a sales colleague and advertise open roles, showing that the account also amplifies Novee Security's team.
For adjacent viewpoints, see Jim Manico on AI-native secure engineering, Deepen Desai on agentic security operations at vendor scale, and Ricardo Camacho on embedded and automotive security verification.
Manico argues that AI is changing the economics of vulnerability discovery and remediation: finding and fixing flaws will become far cheaper, displacing standalone scanners, SAST, DAST and CVE tooling with AI-native capabilities.
Read more
Both accounts pair a strong stance on AI with a defender-side practice recommendation.
Manico focuses on application security engineering and standards: security requirements defined during specification, AI threat modeling, secure code generation, and using OWASP ASVS and AISVS to specify what must be verified. Geffen starts from news events and attacker economics, then makes the case for continuous offensive testing rather than secure-by-specification engineering.
Desai publishes vendor-side commentary on the shift Geffen describes: attackers automating attack stages, chaining exploits and using AI to compress defender response time.
Read more
The posts also cover testing frontier models such as Anthropic Mythos and GPT-5.5 Cyber against real attack and defense scenarios. Both treat agentic security operations as the answer.
Desai's posts are platform and launch communications for a large security vendor, covering unified exposure and threat management, inline telemetry, closed-loop remediation, and named customers and partners. Geffen writes short, source-led arguments for a startup, pressing one thesis: continuous pentesting.
Camacho shares Geffen's premise that reacting after release is too late, arguing that security must be engineered through the product lifecycle.
Read more
They also discuss AI-assisted verification paired with human oversight, and both attach their arguments to longer explanatory resources.
Camacho focuses on embedded and automotive systems: software-defined vehicles, ECUs, ISO/SAE 21434, static and dynamic analysis, fuzz testing, and secure coding in long-lifecycle products. Geffen writes about enterprise defenders facing AI attackers and continuous testing.
For a cybersecurity campaign about AI-driven offensive security, Ido Geffen’s posts show clear subject-matter relevance: they discuss AI-enabled attacks, continuous penetration testing, and Novee Security’s approach.
Read more
Before committing, ask for recent audience data by job, industry, seniority, and location, including documented roles and seniority if targeting security decision-makers.
LinkedIn followers
15,948
LinkedIn profile; follower count recorded in the account data · 2026-10-08
Ido Geffen's posts do not offer a sponsorship programme.
Read more
Across six LinkedIn posts published between April 9 and October 6, 2026, they write as an operator promoting Novee Security rather than as a creator selling placements to outside brands.
The posts cover AI-assisted attacks, continuous penetration testing, and AI cost economics.
They publish no rate card, media kit, audience spec sheet, disclosure of a paid brand collaboration, or partner terms.
The closest thing to an outside-facing editorial arrangement is a signed piece Geffen wrote for CTech by Calcalist on how defenders cannot outspend AI attackers, announced and linked from LinkedIn.
It is a contributed byline in an established outlet, not an advertised sponsorship package, and offers no information a buyer could plan against, such as accepted topics, lead time, pricing, or exclusivity.
The remaining posts are company marketing: launch messaging, product positioning, a sales-award post recognising an account executive, and a recruiting post pointing to Novee hiring.
The recruiting link is the only outward call to action, and it is not a partnership route.
Any arrangement would have to be negotiated directly and built from scratch.
The posts are a small, recent, company-focused sample, so ask for a partnership or media contact, any collaboration policy or disclosure practice, and an example of a past paid or unpaid brand collaboration outside Novee Security.
A cybersecurity vendor selling AI-driven pentesting and seeking a founder-level thought leader to co-create content on continuous offensive testing can work with Ido Geffen.
Read more
The posts support the editorial half of that campaign; the commercial terms have to be negotiated from zero.
The campaign needs someone who can speak credibly about AI offensive security and build a collaboration around it.
Geffen's posts provide that foundation.
First, the problem statement is already theirs: AI writes more code faster and with more flaws while the same AI improves at finding them, and adding security headcount does not close the volume gap, because a team of twenty cannot manually review what an AI coding assistant ships in an afternoon across a thousand engineers.
Second, the argument draws on outside sources rather than marketing language — a New York Times report on the security hiring boom and $7–8M CISO packages, OpenAI's Black Hat evaluation in which agents shared memory and chained exploits into zero-days, a CNBC report, and a reported Anthropic model.
A vendor gets a neutral-sounding setup it does not have to write itself.
Third, the proposed answer is consistent and quotable: continuous, attacker-like testing instead of periodic validation, stated as "not a better scanner, not a bigger model, but a system that runs the way attackers do."
The format suits a collaboration, too.
They write short, structured posts — blunt headline, one named third-party source, a defender-facing conclusion, often a link — and have carried at least one signed off-platform piece, for CTech by Calcalist.
A vendor wanting a co-developed article plus short-form amplification is asking for a format they already produce.
The constraint is commercial rather than editorial.
The posts do not document a sponsorship programme: no rate card, media kit, accepted-topic list, lead time, exclusivity terms, partnership contact, or disclosed paid brand collaboration.
The only outward call to action in the posts here is a Novee hiring link.
Before quoting pricing, deliverables, or timelines, ask for a partnership contact and at least one past collaboration example.
One editorial caution: the recommendation they make — continuous testing — is also Novee's commercial position.
Confirm whether the vendor wants that thesis amplified or its own approach tested against it in the same piece.
When to choose a specialist
Reject this creator for a campaign that depends on documented sponsorship mechanics and audience proof.
Read more
Consider an AI coding-assistant vendor targeting software engineers and engineering managers at mid-size companies, seeking a paid, disclosed sponsored series with published rates, a stated lead time, and audience demographics by role and seniority.
There is topical overlap: one post, the CTech by Calcalist piece, discusses AI coding budgets, Uber burning a year of AI spend in four months, and $1,500 monthly caps per coding tool.
The fit still fails on three counts.
The commercial requirements cannot be verified from the posts.
They do not document accepted formats, pricing, lead time, exclusivity, or an inquiry route, and no paid collaboration outside Novee Security is documented.
A buyer planning a budget and calendar has nothing to plan against.
The audience requirement cannot be substantiated either.
The only audience figure mentioned in the posts is a LinkedIn follower count of 15,948 observed October 8, 2026.
The posts do not establish followers' job, industry, seniority, or location.
Ask for audience demographics by role and seniority before committing to a campaign that depends on reaching engineers.
The framing also runs the wrong way.
Their AI-economics argument is about restraining spend and routing each task to the model built for it rather than renting frontier models for everything.
A vendor whose pitch assumes teams should buy more AI tooling would be borrowing a thesis that points the other direction.
The posts are also security-framed throughout, each one closing on Novee's own position, which is a narrow base for a developer-tooling campaign.
This is a topic and readiness mismatch, not evidence that they refuse sponsorships.
Ask directly whether they accept sponsored series and can provide the required commercial terms.
If the goal is security-adjacent rather than developer tooling, revisit the campaign against the good-fit conditions above.
A worked campaign example
Worked campaign angle for the cybersecurity vendor, where the fit is genuine.
Angle: the defender's cost problem. A co-developed piece arguing that economics, not technology, blocks continuous testing, followed by evidence of what changes when testing actually runs continuously.
Why this angle: it starts from the argument they already make in the Calcalist piece — an attacker needs AI for one target on a cheap, unguarded open-source model, while a defender must test everything continuously, mostly on frontier models — and gives the vendor a contribution missing from their post: its own data on inference cost per test run, how tasks are routed to cheaper models, and what continuous validation catches that a periodic scan misses.
Format, matched to the posts:
One signed long-form article shaped like the Calcalist byline: one named external source, one quotable cost figure, one architectural claim.
One short LinkedIn post in the format they use: blunt headline, a single third-party anchor, four to six paragraphs, one closing link.
One recorded technical conversation in which the vendor's engineers demonstrate continuous testing against a known vulnerability class while they supply the attacker-side framing.
Sequence: 1. Problem. AI writes more code while the same AI improves at exploiting it, and more security hires do not close the volume gap. Anchors: the New York Times report on the security hiring boom and OpenAI's Black Hat evaluation. 2. Cost. Defender economics — testing everything continuously versus an attacker testing one target cheaply. Anchors: vendor-provided cost data plus the Uber budget example. 3. Demonstration. What continuous validation finds in a week that periodic validation missed.
Before finalising, ask for a partnership contact, accepted topics, lead time, exclusivity terms, and at least one prior collaboration example. If the campaign depends on reaching security decision-makers, request audience demographics by role and seniority. Without that information, the collaboration can be proposed but not scheduled, priced, or promised.
Decision point for the vendor: state up front whether it wants their continuous-testing thesis amplified or an independent comparison against its own approach. Both are workable; leaving it unstated is what produces a piece nobody can approve.
Choose Manico when the collaboration should be a standards-and-economics argument aimed at application security teams.
Read more
The two posts argue that AI agents can increasingly analyze whole codebases, discover vulnerabilities, develop exploits, generate patches, write regression tests, and verify results, changing the economics of both finding and fixing flaws.
From that, the prediction is that stand-alone scanners, threat-modeling tools, SAST, DAST, IAST, and CVE scanners get absorbed into AI-native capability, while OWASP ASVS and AISVS matter more because they define what security means and what must be verified.
Human judgment is reserved for architecture, business logic, and novel attacks.
A vendor selling verification or remediation rather than detection gets an argument about where value moves next, tied to standards a buyer's engineering team may already use, plus a live example: agent swarms attacking RubyGems, hundreds of packages involved, signups paused.
Manico interprets the shift rather than operating it.
Ido Geffen's posts describe what AI agents can do and what should replace existing tooling; they do not describe a product, a running pentesting programme, or an incident they handled.
A vendor recruiting a founder-level builder for an AI-pentesting collaboration gets a credible analyst of the change rather than someone with a live system to demonstrate.
The position is also an argument: one post defends the AI-cyber thesis against doubters and asks people with daily frontier-model experience to push back if they disagree.
That is useful when the vendor wants a debate it can host, and awkward when it wants a steady product narrative.
Choose Desai when the vendor wants a platform-scale voice that has already run frontier models against real attack and defence scenarios and can speak about the operating model rather than the thesis.
Read more
The posts describe an agentic security operations launch built on unified exposure and threat management, inline telemetry across network, identity, endpoint, cloud, AI, and deception, specialized AI agents shaped by frontline SOC and threat-hunting experience, and closed-loop remediation that turns detection into action.
A second post describes testing Anthropic's Mythos and OpenAI's GPT 5.5 Cyber on live scenarios and reports three observations: multi-step reasoning lets a model build complex exploit paths rather than scan for known CVEs, architectural context improves exploit accuracy, and context should only ground the model to its environment.
For a vendor whose pitch is integrated detection and remediation rather than a stand-alone scanner, that vocabulary is closer to the buyer's.
The framing is corporate.
The posts sit inside one company's platform launch and name its launch partners and customers, so a co-developed piece would be measured against an existing vendor narrative, and a smaller company risks reading as the junior partner in someone else's announcement.
They also stay at the level of capability and platform architecture.
A buyer seeking a specific customer engagement, reusable cost or deployment detail, or a view of how Desai works outside the company's communications should ask for those details.
Ido Geffen offers an independent founder thesis about attacker economics and continuous testing.
Choose Camacho when the buyer is a product engineering team rather than an enterprise security operations centre, and the deliverable is testing methodology rather than an AI-attacker race.
Read more
One post argues that software-defined vehicles are defined by software—millions of lines of code, hundreds of interconnected ECUs, over-the-air updates, ADAS, AI-powered features—and that telematics systems, ECU communications, OTA mechanisms, and vehicle lighting have all been exploited.
It frames ISO/SAE 21434 as what turns cybersecurity into an engineering discipline and puts AI-assisted verification alongside human oversight.
The second post lays out the three approaches Camacho recommends: static analysis, dynamic analysis and fuzz testing, and penetration testing, plus shifting security left, secure coding, CI/CD security gates, and defence in depth.
A vendor selling verification tools into automotive or embedded engineering gets a ready-made audience and a standards frame it can attach to.
The scope is embedded and automotive, and the content is instructional.
The posts teach engineering teams how to structure testing and what standards require; they make no first-hand claim about AI offensive work, incident response, or AI-driven pentesting as a product.
For the same campaign decision—an AI-pentesting vendor recruiting a founder-level voice—Ido Geffen supplies the attacker-side economics, the Black Hat agent-collaboration example, and the continuous-testing argument, while Camacho supplies a domain methodology that would have to be re-argued for a different audience.
If the collaboration is with an automotive or embedded vendor instead, the tradeoff reverses and Camacho becomes the stronger starting point.
Ido Geffen's LinkedIn posts include six full texts, not just headlines or excerpts: AI security arguments anchored to OpenAI's Black Hat talk, New York Times reporting, CNBC coverage, and a CTech by Calcalist byline, plus a sales-recognition post and a recruiting post.
Comparisons with similar creators draw on two full posts each from three other accounts.
Geffen's posts do not state their gender, so this profile repeats the surname or uses they.
No individual publication dates appear for Geffen's posts, so no posting cadence is claimed, and the Calcalist article itself appears here only through Geffen's own summary.
No information on signature books, shows, or earlier career work appears here, and no sponsorship, rate-card, or audience-demographic detail appears in the posts here; those remain open gaps.
