Argues the real story isn't the geopolitics but the supply chain: BlackCore represents the maturation of influence ops from state capability to bespoke contractor work to a productized vendor market with SLAs, account inventory, and reseller channels. Compares the trajectory directly to NSO Group's spyware productization arc and frames BlackCore as a successor to Archimedes Group and 'Team Jorge.'
Reports that BlackCore is suspected of running coordinated inauthentic accounts across X, Facebook, TikTok and Telegram, deploying AI-generated avatars and voice clips of named opponents, and operating bespoke astroturf websites in the NYC mayoral race, a Scottish vote, and a French election. Frames these as turnkey campaigns sold to political clients for opposition smear, sentiment shaping, and GOTV suppression.
By submitting the Reuters story and driving it to 168 points, surfaces the allegation that an Israeli private-sector firm is actively interfering in Western elections as a topic worth developer attention.
Notes that neither Meta nor X confirmed any takedowns tied to BlackCore despite being contacted, that the Israeli government declined to comment, and that the firm itself denies the allegations. Implicitly frames this as a pattern of non-response that lets the vendor market keep operating.
Reuters reported on June 11 that BlackCore, an Israeli private-sector influence-operations firm, is suspected of covertly meddling in the New York City mayoral race, a Scottish vote, and a French election. The story landed at 168 points on Hacker News, which is unusual — election-meddling pieces normally die in /new. This one didn't, because the comment thread quickly pivoted from politics to the part developers actually care about: BlackCore appears to be a productized successor to Archimedes Group and 'Team Jorge,' the Israeli outfits Facebook took down in 2019 and that Forbidden Stories exposed in 2023 — and the productization is the story.
The specific allegations, per Reuters' reporting and prior work by DisinfoLab and Graphika on related networks: coordinated inauthentic accounts seeded across X, Facebook, TikTok and Telegram; AI-generated avatars and voice clips attributed to named opponents; targeted comment-spam under local-news posts; and bespoke 'astroturf' websites that get cited back into the laundering loop. The operations were sold to political clients as turnkey campaigns — opposition smear, sentiment shaping, get-out-the-vote suppression — priced and scoped the way you'd scope a brand campaign.
None of the platforms have publicly confirmed takedowns tied to BlackCore as of the report. Meta and X were both contacted by Reuters; both gave the usual non-answer. The Israeli government has not commented. The firm itself, predictably, denies everything.
The first-order story is geopolitics. The second-order story, and the one HN actually engaged with, is the supply chain. Influence operations have followed the exact arc that offensive security did a decade ago: from in-house state capability, to bespoke contractor work, to a productized vendor market with SLAs, account inventory, and reseller channels. NSO Group did this for spyware. BlackCore — and at least four other named Israeli, UAE-based, and Spain-based firms tracked by DisinfoLab — are doing it for narrative.
What does 'productized' mean concretely? Three things. First, persona inventory at scale: aged accounts on every major platform, warmed with months of plausible activity, regionally diverse, ready to deploy. The hard part of an influence op is no longer writing the content — LLMs do that — it's having credible-looking accounts to post from that survive the platforms' 'coordinated inauthentic behavior' classifiers. That inventory is now a line item you buy. Second, platform-evasion R&D: residential proxies, mobile-carrier IP rotation, device fingerprint randomization, behavioral pacing models tuned against each platform's known heuristics. Third, measurement: dashboards showing reach, sentiment lift, search-trend movement on the target candidate's name. You can A/B test a smear.
The HN thread surfaced the part the Reuters piece soft-pedals: the platforms' integrity teams have been gutted since 2023, and the vendors know it. X dissolved most of its election-integrity work; Meta shrank its civic-integrity team during the layoff rounds; YouTube quietly walked back its 2020-election misinformation policy in 2023. The detection asymmetry is now extreme. One senior commenter who'd worked T&S at a FAANG put it bluntly: 'We used to catch Archimedes-class ops in days. We caught them because we were staffed for it. That staffing is gone. The ops are not.'
Third-order: this is now a regulatory inevitability. The EU's DSA already obligates VLOPs to assess and mitigate 'systemic risks' from inauthentic coordinated behavior; expect BlackCore to be cited in the next round of DSA enforcement letters. The UK's Online Safety Act has a narrower hook but Ofcom is reportedly already reviewing the Scottish-vote allegations. In the US, there is no federal framework, which is precisely why NYC was a target — local races have no integrity apparatus at all.
If you run anything that accepts user-generated content with civic adjacency — a news comment system, a forum, a polling widget, a reviews site, even a Slack-style community — BlackCore is your threat model now, whether or not anyone has bought your platform yet. Three concrete implications.
First, your bot-detection budget needs to assume the adversary has dedicated R&D against your stack, not a script kiddie running puppeteer. reCAPTCHA v3, hCaptcha, and basic device-fingerprinting are checkbox compliance. They will not stop a vendor selling residential-IP'd, aged, behaviorally-paced accounts. The defenses that actually work — graph analysis on posting cadence, cross-account stylometric clustering, network-level correlation across content — are not off-the-shelf. Budget for either building this in-house or buying from vendors who actually do graph-level detection (Spectrum Labs, Cinder, Hive's coordination models). 'We use Cloudflare bot management' is not a serious answer in 2026.
Second, provenance is now table stakes for any user-generated content workflow. C2PA content credentials, model-watermark verification (for what little of it survives a screenshot), and origin metadata need to be in your ingestion path, not a 2027 roadmap item. If you're shipping any feature that lets users post media into a public feed, the question 'can a reader cheaply verify this wasn't generated five minutes ago by a persona farm' has to have an answer.
Third, if you sell SaaS to political campaigns, civic orgs, or news publishers, you are going to start getting procurement questions about your coordinated-inauthenticity controls. Have the SOC 2-style answer ready. 'We rely on platform-level moderation' will fail the question.
The NSO Group analogy is the right one to hold in your head. Spyware went from 'one bespoke tool used by one agency' to 'commodity sold to 40+ governments and several private clients' in under a decade, and the regulatory response is still catching up. Influence-ops vendors are five years into the same curve and accelerating, because LLMs collapsed the content-generation cost to zero and platform layoffs collapsed the detection cost to near-infinite. BlackCore is the name on this week's story. There are at least a dozen others, and the next NYC-class race is six months away.
I confused BlackCore with Black Cube, a different Israeli private oppo research and dirty tricks group of former intelligence agents. They gained attention for their dirty campaigns against Harvey Weinstein's accusers, NSOs critics and Hungarian opposition.https://en.wikipedia.org
"Lecornu said the French government had asked Israel for explanations of BlackCore's actions, and also for help in trying to find out who may have been behind the smear campaign."This is a very well executed bit of diplomacy.
The israeli ambassador in France should already have been kicked out a while ago for a myriad of reasons, I'm ashamed my country is so spineless.
Is this the same company that Slovenia was asking the EU for help with regarding the company's meddling in the election process?
Top 10 dev stories every morning at 8am UTC. AI-curated. Retro terminal HTML email.
As a New Yorker this doesn’t shock me too much. The level of “Mamdani is an anti-Semite” sentiment I saw online (Reddit particularly) felt truly hysterical. And wasn’t matched by any equivalent in the offline world.