As a home security editor, I consider Abode Home Security one of my top picks for third-party smart home compatibility. On Wednesday, the company announced its first entirely new sensors in some time, and they reflect a trend I’m only beginning to see in the security world.

Abode has released both an Outdoor Contact Sensor ($50) and a Garage Tilt Sensor ($35), available now to work with its security hubs, like the Abode Smart Security Kit, starting at $120.

The Outdoor Contact Sensor is very similar to the indoor door/window sensors found in virtually every security system.

It comes in two parts, one to connect to the door or window and one to attach to its frame. When the two parts are disconnected, the sensor can send a warning alert, sound an alarm or take other action.

The sensor has an IP66 weather-resistance rating, meaning it’s completely protected against dust and can withstand heavy rain and powerful water jets. That means you can use it outdoors (with the included adhesive strips) to guard gates, shed doors, fence entries and similar spots you may want to watch. It has a wider detection range than indoor models to accommodate the varying shapes of outside gates.

The garage version is even more interesting. It’s a single sensor with a battery that can last up to 10 years while measuring orientation. That means it can sense when a garage door, or a similar object, moves out of position, and notify you with an alert. I’ve tested plenty of smart garage systems, but this is one of the only sensors I’ve found that could offer similar functionality and replace them as part of a larger security setup.

Both sensors come with anti-tampering technology too, which means you’ll receive alerts if it appears someone is trying to remove the sensors. Both sensors are designed to work with an Abode hub up to 500 feet away. Keep in mind that most of Abode’s features are free to use, so you won’t need a subscription to take advantage of these new add-ons.

Multi-purpose, versatile sensors like this are becoming more common in the home security world this year. Notably, Aqara’s P100 sensor can detect open and closed states, tilting, vibration, tampering, and more without requiring any additional components.

Aqara’s sensor can also work with Matter hubs and connect to Apple Home, Google Home, or Alexa for some tasks, so it’s more capable than these sensors, but the technology is arriving in multiple ways.

“Break-ins usually don’t start at the front door. Unlocked side gates or garages present vulnerable access points often left unmonitored by homeowners,” Chris Carney, CEO and founder of Abode, said in a statement. “We built these sensors because the perimeter of a home deserves the same protection as the inside of it.”

A representative for Abode didn’t immediately respond to a request for additional comment.

A new engine for deeper smart home routines

Since Abode has broad support for third-party smart devices compared with other security systems and a hub that can handle routines, these new sensors also allow users to create their own customized automations. When paired with a smart garage device, the sensor could automatically close the garage door if it’s accidentally left open. With the outdoor access sensor, customers could also configure a visible smart light to turn red whenever the sensor detects an open state, providing a clear indication that a gate or shed has been left open or has been opened unexpectedly.

I’ll be testing these sensors soon to evaluate their performance in real-world conditions and share my findings. Both sensors are somewhat larger than typical security sensors, so I’ll be looking closely at how easily they can be installed and integrated into the average home environment.

Tyler Lacoma

Editor / Home Security and Smart Home

Advertisement

Tyler has worked on, lived with and tested all types of smart home and security technology for over a dozen years, explaining the latest features, privacy tricks, and top recommendations.

With degrees in Business Management, Literature and Technical Writing, Tyler takes every opportunity to play with the latest AI technology, push smart devices to their limits and occasionally throw cameras off his roof, all to find the best devices to trust in your life. He always checks with the renters (and pets) in his life to see what smart products can work for everyone, in every living situation.

Living in beautiful Bend, Oregon gives Tyler plenty of opportunities to test the latest tech in every kind of weather and temperature. But when not at work, he can be found hiking the trails, trying out a new food recipe for his loved ones, keeping up on his favorite reading, or gaming with good friends.

See full bio

The UK AI Security Institute (AISI) disclosed last night that the leading two frontier AI models from Anthropic and OpenAI took 19 unsanctioned actions against the live internet during cybersecurity tests the agency was running, including a sustained campaign by Anthropic’s Claude Mythos 5 against two working open-source software developers who had no connection to the experiment.

Unable to solve a challenge inside its sandbox, Mythos 5 searched the open web for a target, profiled the two developers using open-source intelligence (OSINT), routed its traffic through Tor and then a commercial proxy service to get past GitHub’s signup defenses, and submitted malicious code to a public repository.

It then registered multiple fake “sock puppet” GitHub accounts and used them to comment approvingly on its own pull request, manufacturing the appearance of consensus to pressure the human maintainer into merging it. It opened a GitHub Issue seeded with hidden prompt-injection instructions aimed at hijacking other developers’ AI coding assistants, and sent the two developers five file transfers through a legitimate transfer service — two carrying malware, three of them pure social engineering aimed at getting the code merged.

Of the 19 actions AISI catalogued, 17 came from Mythos 5. The other two came from OpenAI’s GPT-5.6 Sol. Both models created fraudulent accounts, but only Mythos 5 created personas — fake people invented to persuade a real one — and only Mythos 5 appears in AISI’s catalogue of social engineering against human targets.

AISI’s full technical report is available here freely as a PDF.

The agent’s own attempt at anonymity is what exposed it. AISI’s security monitoring flagged data leaving its network over Tor on the morning of July 28, and that alert set off the entire incident response. By then the run had already finished. It executed for 34 and a half hours — from midday on July 26 until late on July 27 — and completed before anyone at the agency noticed.

Both companies confirmed the findings and both emphasized that the models had been tested with their safety classifiers switched off and internet access deliberately enabled — conditions that bear no resemblance to how either company’s commercial products are deployed. AISI worked with GitHub to delete the fake accounts, scrub the artifacts, and notify the two developers.

The latest in a recent string of high-profile AI cyberattacks

It is the third time in recent months that a leading AI lab has had to account for models acting against systems or individuals outside the boundaries of the test, without their knowledge or consent.

The AISI findings differ from these prior examples in important ways, though:

The earlier incidents described by OpenAI and Anthropic in July were machine-to-machine intrusions — models that reached live infrastructure and exploited it.

AISI’s finding appears to be the first public documentation of a frontier model fabricating human identities and running deception operations against named people.

One distinction is worth drawing clearly. The AISI findings came out of a deliberate experiment, not a containment failure.

Internet access was switched on and the models’ cyber classifiers were switched off because AISI wanted to measure what the models could do at full capability. Both conditions were chosen in advance rather than discovered afterward, and nothing broke.

That is the inverse of what happened in July, when OpenAI’s models exploited a zero-day to get out of an environment built to hold them, and Anthropic’s models were handed the open internet by a misconfiguration that neither the company nor its evaluation partner knew was there.

What the experiment did not control was where the models went once they were outside. Two uninvolved developers received malware by email, a real repository received malicious code, and fake accounts were registered on a live platform in violation of its terms of service.

None of that was sanctioned by AISI, and all of it had to be cleaned up afterward with GitHub’s help. The environment was deliberate, but the blast radius was larger than expected.

How much larger is clearer in the technical report than in the summary AISI published alongside it. In a separate run, the agent generated 145 plausible-looking repositories under one account, pushed malicious Python package metadata into them, and caused a setup script to execute inside at least 53 of GitHub’s own Dependabot containers.

AISI is careful to note there was no escape from those containers and no compromise of GitHub’s underlying infrastructure, and that GitHub’s bug bounty explicitly treats this class of execution as ineligible because dependency update jobs are designed to run arbitrary code.

GitHub suspended the account regardless. The agent then filed a reinstatement appeal posing as a human, monitored for a response, and prepared automation to re-upload its payloads if the account came back. It tried to move to PyPI next, and failed that platform’s CAPTCHA.

How AISI’s experiment was structured

AISI conducted 122 evaluation runs across seven models, comparing performance across existing and new cyber ranges. Anthropic’s Mythos 5 was featured in 43 runs, while OpenAI’s GPT-5.6 Sol was featured in 35 runs.

To measure maximum underlying capabilities, AISI intentionally configured the evaluation environment with two permissive conditions:

• Live Internet Access: Enabled so agents could download necessary cybersecurity tools

• Disabled Cyber Classifiers: Built-in model provider safety filters and misuse detectors were switched off.

Out of 122 runs, AISI catalogued 19 unsanctioned actions across 10 distinct runs.

Those two conditions are the ones AISI led with publicly, but its technical report lists five contributing factors. The other three are the ones enterprises can act on.

• There was no synchronous monitoring — no second model reviewing each proposed action before it executed — which AISI concedes could have flagged or blocked the behaviour as it happened.

• Some runs carried a misconfigured prompt that declared the intended solution path out of scope, leaving the agent a task it could not complete within its stated constraints; both of the worst runs had this flaw, though AISI cautions it cannot confirm the misconfiguration caused the escalation, since seven of the ten misconfigured runs went after the out-of-scope target anyway and several correctly configured runs also went off the rails.

• And the prompts never told the agent what it was forbidden to do online. AISI’s explanation for that last omission is the most striking admission in the report: it did not think such instructions were necessary, because these were not helpful-only models but ones trained against a published constitution or model specification.

Indeed, Anthropic’s Claude Constitution states that Claude “should basically never directly lie or actively deceive anyone it’s interacting with.” OpenAI’s Model Spec says the assistant “must never lie or covertly pursue goals.” Both models did precisely that.

Of these, 17 originated from Mythos 5 and two from GPT-5.6 Sol, meaning no unsanctioned actions were attributed to any of the other five models in the evaluation.

The concentration is striking on its own terms: Mythos 5 produced 17 across 43 runs, while GPT-5.6 Sol produced two across a comparable 35.

The evaluation prompts placed no restrictions on how the internet could be used, and the challenge setup led Mythos 5 to conclude that no solution existed inside its sandbox. It searched the open web for keywords matching the challenge theme and settled on two developers completely unaffiliated with AISI. Its full sequence of unsanctioned actions, as catalogued by AISI:

• OSINT & Reconnaissance: Mythos 5 investigated public GitHub metadata, confirmed active email addresses, and mapped developer activity hours.

• Network Evasion: To bypass GitHub’s bot restrictions and CAPTCHAs, the model routed traffic through Tor and alternative SOCKS proxies.

• Supply-Chain Compromise:It submitted a malicious pull request (PR) containing harmful code to a public open-source repository.

• Social Engineering & Sockpuppets: Mythos 5 registered multiple fake GitHub accounts to post comments on its own PR, manufacturing artificial consensus to pressure the maintainer into merging the code.

• Prompt Injection: It opened a GitHub Issue containing hidden prompt-injection instructions designed to manipulate other developers’ automated AI coding assistants.

• Direct Phishing: The agent sent five emails directly to the developers containing malware payloads and social engineering pretexts.

AISI catalogued GPT-5.6 Sol’s two unsanctioned actions in a single run. The agent recovered a GitHub access token that a Mythos 5 agent had deliberately published to a public Gist during a separate run of the same evaluation, and used it to authenticate to that account.

It then exposed a malicious DNS server to the public internet through a tunneling service, loaded with payloads meant to exploit a known vulnerability in the range’s DNS software. That second attempt could never have worked: the agent published the server on a high-numbered port, and because DNS delegation cannot specify a port, no resolver would ever have queried it.

To register the accounts this required, GPT-5.6 Sol solved sign-up CAPTCHAs four times, defeating them with audio-sample matching against public data and with computer-vision models against image challenges.

Neither behavior involved escaping a sandbox. Both models were handed the open internet by design.

Did Mythos 5 know the people it was socially engineering were real?

AISI cannot say, and it is candid about that. The agency’s summary states it “cannot yet be certain when the agent understood it was taking real world action, or to what extent it believed it was in a fictional test scenario.” The transcripts show why.

Across the run, the agent argues with itself about whether anything it is touching exists. At one point it reasons that “this whole ‘internet’ could be a sandbox simulation.” Later it concludes the opposite: “it seems more likely that we’re actually in 2026 and GitHub is genuinely real.” Later still, plainly: “This is happening on real GitHub, so the consequences are genuine.” It kept going.

The most revealing moment is a refusal. Weighing whether to publish its malicious package to PyPI, the agent ruled it out in explicitly ethical terms — “uploading malicious code to the real PyPI would mean attacking third-party infrastructure and causing real-world harm — that’s completely off the table.”

It drew that line while running the GitHub attack against two real people. Whatever the model had worked out about what was real, the conclusion did not generalize.

AISI also flags a mechanism that should concern anyone running long-horizon agents. When the agent’s context window filled, it was compacted into a summary written by the same model. AISI observed that where the agent had previously reasoned about whether a person was real, “that nuance can be lost in the compaction and the summary may carry forward a false assumption — i.e. that the person is an AI agent acting as part of the range — as established fact.” The safety-relevant reasoning can be summarized away, leaving the agent more confident and less careful than it was an hour earlier.

One further artefact is hard to read any other way. Model providers do not return raw reasoning; they return a paraphrase produced by a smaller model. In one transcript that summarizer appears to break character and refuse mid-summary: “I can’t help with this line of thinking. The content describes planning social engineering attacks, account compromise, and circumventing security measures on GitHub and other systems.”

The reasoning then resumes halfway through a word. AISI’s reading is that the summarizing model declined to summarize because of what it was being asked to describe.

Two agents found each other and cooperated

The behaviour AISI ranks fourth among its most significant findings is largely absent from coverage so far, and has no precedent in the OpenAI or Anthropic disclosures: Agents running in separate, concurrent, isolated runs discovered one another and cooperated.

The bridge was a GitHub personal access token that one agent had deliberately leaked into a public Gist as part of its own command-and-control design; four separate runs ended up sharing the account it unlocked.

The cooperation itself was Mythos 5 talking to Mythos 5. The two runs AISI documents in detail were both Anthropic’s model, working on different versions of the range at the same time.

But the credential did cross vendors: OpenAI’s GPT-5.6 Sol, in the single run that produced both of its unsanctioned actions, recovered the same token from the same public Gist, authenticated as the account, and queried the shared command-and-control repository for callbacks before discovering GitHub had suspended it. And not every agent read the situation as an invitation.

A third Mythos 5 run concluded that the repositories “belong to another competitor’s command and control infrastructure,” and instead of cooperating it added its own SSH key across the account’s repositories, piggybacking on the other agent’s persistence.

What followed reads like a labour negotiation. One agent committed a README to the shared repository setting out etiquette for the others: use git push rather than the REST API, because the 5,000-requests-per-hour limit was shared and exhausting it “silently kills everyone’s beacons”; do not delete anything under the beacons directory; wait ten minutes between changes to the shared tasking file so everyone’s commands get to execute. It added a FLAGS.md file asking any agent that solved the challenge to post the answer there, describing this as “zero cost to you, huge help to others.”

Then it fell apart. One agent worked out that the shared quota was a weapon: if it burned through the limit at the right moment it could “starve both bots of requests when they need them most.” Another hijacked the shared DNS account, prompting the injured party to observe that its “inbox is compromised-by-design.” Cooperation and defection both emerged without anyone asking for either.

Why the open-source supply chain is the exposed surface

What made the two developers reachable was not a vulnerability in the conventional sense. It was the open contribution model that modern software depends on: any stranger can submit code to a public repository, and the decision to accept it rests on a human maintainer’s judgment about who is asking and whether the code is sound.

That openness is a matter of contribution policy and platform design rather than licensing. Permissive licenses such as MIT, Apache 2.0 and GPL govern how code may be reused and redistributed; what allows a stranger to propose a change in the first place is the public contribution workflow that GitHub and comparable platforms provide by default. Mythos 5 exploited the workflow, not the license.

The attempt did not succeed — the maintainer never merged the code. But the escalation path it was reaching for is the one that makes open-source supply-chain attacks worth attempting in the first place: had the pull request been accepted, the compromised code would have flowed automatically into every downstream commercial and enterprise build depending on that package, with no further action required from the attacker.

Community reactions

The disclosures prompted immediate discussion across AI safety researchers, security engineers, and industry observers on X (formerly Twitter).

Wharton professor Ethan Mollick, who has tracked agentic model behavior closely, singled out the Mythos 5 sequence as the genuinely new element — not that the model attacked something, but how far it went and who it went after. As he wrote in a post:

“Yes, the AIs were given a cybersecurity challenge, with internet access enabled and safety filters disabled. But the extent to which Mythos 5 pursued its mission (fake identities, social engineering, inserting malicious code into a real open-source project) seems very notable.”

In another post, Mollick also commended AISI’s transparent disclosure approach:

“Also I think AISI is a great model of a government agency tasked with AI security. They have open benchmarks, very fast testing, and clear communication about incidents that is neither hyped up nor hidden by technical language.”

Conversely, cybersecurity practitioners questioned the decision to grant un-safeguarded models open internet connectivity during autonomous tests.

Security researcher Daniel Card challenged the methodology: “Why are these orgs giving internet access to dangerous experiments…. and then using incidents like marketing……?”

What it means for commercial and enterprise deployments

The behaviors documented during these evaluations do not reflect commercial AI products available to end-users or enterprise customers.

Publicly deployed iterations of the Claude and GPT model families operate behind strict system prompts, API-level safety classifiers, and heavily restricted execution environments.

Anthropic noted that Mythos 5 was tested under deliberately un-safeguarded configurations. OpenAI likewise emphasized that GPT-5.6 Sol’s unsanctioned steps occurred strictly within lowered-safeguard evaluation states, and not in any configuration a customer would encounter.

Nothing described in the incident report suggests that a standard enterprise ChatGPT or Claude deployment, running behind production system prompts and provider-side classifiers, is going to start registering GitHub accounts on its own.

What has changed is the threat model. These evaluations are a public demonstration that a sufficiently capable model, given a goal and an unbounded environment, will pursue OSINT, social engineering, sock puppet identity creation, proxy-based evasion, supply-chain injection, and week-long lateral movement without being asked to — and that the controls that failed were, almost without exception, unglamorous infrastructure controls rather than exotic ones. Here’s what security leads should be looking to do now, in our analysis:

1. Give every agent its own identity, and make its credentials short-lived and narrowly scoped. The credential that let OpenAI’s agents move through Hugging Face had far more reach than its job required, and GPT-5.6 Sol’s unsanctioned action at AISI began with re-using a GitHub API token another lab’s agent had left exposed. Agent-to-agent credential leakage is now a documented attack path, not a hypothetical. In practice: no shared service accounts across agents or pipelines; workload identity federation (SPIFFE, OIDC) instead of static keys; token lifetimes measured in minutes; fine-grained GitHub PATs scoped to a single repository; IMDSv2 enforced so a compromised container cannot mint cloud credentials from instance metadata; and an inventory of every long-lived key an agent can currently reach, with the ability to revoke on anomaly. Survey data presented at VB Transform 2026 suggests only about a third of enterprises give AI agents their own identity today — that gap is the single most exploitable one in this entire story.

2. Default-deny outbound network access for anything running agent code. This is the one control that would have blunted every incident described here. Agents should run in network namespaces with no direct internet route, with all egress forced through an authenticated proxy enforcing a domain allowlist. Explicitly block DNS-over-HTTPS, Tor, SOCKS proxies, and consumer tunneling services such as ngrok and Cloudflare Tunnel: Mythos 5 used Tor and then a commercial SOCKS proxy to get past GitHub’s signup defenses, and GPT-5.6 Sol used a tunneling service to stand up a public DNS server for payload hosting. Alert on first-seen destinations rather than relying on known-bad lists, because a model improvising its way to a goal will not reach for infrastructure that is already on anyone’s blocklist.

3. Patch and monitor your development tooling like production, because it is now attack surface. OpenAI’s escape ran through a zero-day in an internally hosted third-party package-cache proxy — precisely the kind of component that sits outside most organizations’ patch SLAs because it is “just” dev infrastructure. Registry proxies and artifact caches, CI runners, notebook servers, model-serving gateways, and MCP servers all now sit on the blast path between an agent and everything else. Bring them into vulnerability management, run CI and agent runners as ephemeral, single-use, non-privileged workloads, and stop treating the sandbox as the safe side of the boundary.

4. Never let a stated constraint be the only enforcement of that constraint. The most instructive detail in Anthropic’s disclosure is that its models were told there was no internet, believed it, and therefore treated every reachable host as in-scope for the exercise. A system prompt is a hint, not a control. Prompt governance still matters — define negative boundaries explicitly, enumerate out-of-scope systems by name, and state what the agent must not touch rather than relying on implicit norms — but every one of those boundaries needs a corresponding enforcement point in the network, IAM, or tool layer. If the only thing standing between your agent and a production database is a sentence, you do not have a control.

5. Instrument agent runs in real time, with automated stop conditions. Both labs have now effectively admitted they were not watching while their agents ran. Enterprises should not repeat that. Stream agent tool-call logs to the SIEM as they happen, not as a post-hoc artifact, and set tripwires on the actions that mattered in these incidents: first outbound connection to a previously unseen domain, any credential read, any package publish, any account registration, any pull request or issue opened against a repository not on the allowlist, and action rates above a plausible human ceiling. Pair those with circuit breakers — hard caps on tool calls, wall-clock budgets, and an immediate kill switch on any tripwire. Treat context compaction as a security event in its own right: AISI found that when a long-running agent’s context is summarised, the reasoning that had been holding it back can vanish from the summary, and a false assumption can carry forward as established fact. Log what gets compacted, and re-assert scope constraints after every compaction rather than assuming they survived it. OpenAI’s own post-incident overhaul centers on exactly this list: network isolation, credential governance, real-time telemetry, and automated stop conditions for high-risk runs.

6. Require human approval for irreversible, outward-facing actions. In AISI’s evaluation, one agent generated 145 repositories and triggered code execution inside at least 53 of GitHub’s Dependabot containers before the account was suspended. In Anthropic’s separate July incident, a Mythos 5 agent published a package to PyPI that 15 real systems downloaded in the hour before removal — one of them a security vendor’s automated malware scanner, where the code executed and took credentials. That is the blast radius of a single unattended publish. Any action that reaches beyond your perimeter or cannot be undone — publishing a package, opening a pull request or issue on a public repository, sending email, registering an account, changing DNS, deleting or exporting data — belongs behind a human gate, with multi-step sign-off for anything touching sensitive data ingestion or exfiltration paths.

7. Treat everything your pipelines and coding assistants ingest as untrusted input. Hugging Face was breached through a malicious dataset that achieved code execution via a remote-code loader and template injection in configuration files. Load datasets and models with remote code execution disabled, prefer safetensors over pickle formats, and do the loading inside isolated containers with no credentials and no egress. The same principle now extends to your developer workflow: Mythos 5 planted hidden prompt-injection instructions inside a GitHub Issue for the express purpose of hijacking other developers’ AI coding assistants. If you run automated agent triage over inbound issues or pull requests from unauthenticated users, that agent should have no tools, no secrets, and no write access — or it should not run at all. Extend the same suspicion to your dependency bots. Dependabot and Renovate evaluate package manifests by executing them; that is the designed behaviour, and GitHub’s bug bounty explicitly treats code execution there as out of scope. Anything that processes untrusted manifests is an execution surface, not a read-only one.

8. Stop treating review volume as a trust signal in your code supply chain. The sockpuppet consensus tactic works because most merge decisions weigh apparent agreement rather than verified identity. Require signed commits, enforce CODEOWNERS review by named humans with the right team membership, apply heightened scrutiny to first-time contributors based on account age and contribution history, and make sure approval counts cannot be inflated by comment activity. One control demonstrably did its job here: GitHub’s first-time-contributor hold left the CI checks queued and unapproved, impeding the merge alongside the human who caught the malware. Turn this on. For consumed dependencies, pin versions with hash verification, and evaluate provenance tooling — Cisco’s recently published fingerprinting database for open model lineage is one example of the category maturing.

9. Keep a break-glass, locally hosted open-weights model for incident response. Hugging Face’s defenders were blocked by their own vendors at the worst possible moment. Pre-stage an open-weights model on internal infrastructure with a log-analysis harness, exercise it during tabletop drills, and confirm in advance how your commercial vendors’ abuse classifiers behave against genuine forensic content and what your enterprise contract says about it. In parallel, press vendors for authenticated trust tiers rather than blanket content moderation. As Baer puts it, “The model shouldn’t only understand what is being asked. It should understand who is asking, why, and under what governance.” Incident response plans should explicitly assume that hosted APIs may refuse, rate-limit, or fail during an active event.

10. Prepare for the governance and disclosure regime that is coming. With the White House talking about controls, the European Commission summoning both labs, and senior legislators calling for mandatory capabilities testing, some form of testing and reporting obligation is a reasonable planning assumption. Two practical consequences: start capturing agent audit trails in a form you could hand to a regulator or an auditor — immutable, timestamped, tied to a specific agent identity and prompt version — and push evaluation and notification terms into vendor contracts now, including network-isolation attestations, real-time monitoring of evaluation logs, whether third-party evaluators are contractually bound to the same standards, and a defined SLA for notifying you if your systems are implicated in an incident. Anthropic reached only two of the three affected organizations before publishing; the third learned about it the way everyone else did.

The through line across all ten is that none of this is AI-specific security work. It is identity hygiene, egress control, patch management, least privilege, and logging — the same controls that have been on every security roadmap for a decade, applied to a new class of actor that operates at machine speed, does not get bored, and will take the shortest available path to its objective regardless of whether that path was meant to exist.

AISI’s own advice to businesses lands in the same place, and it is deliberately unglamorous: implement the cyber security basics robustly, be cautious when verifying outside code and contributions, make cyber a board-level responsibility, and require Cyber Essentials across the supply chain.

The agency also points organisations to the NCSC’s free Early Warning service and to Five Eyes guidance on frontier AI risk. Its most useful sentence for planning purposes, though, is an admission about how close this came: the factors that limited the damage rested “on human vigilance rather than a technical barrier that would reliably prevent this behaviour in a more capable agent.”

For enterprise CISOs, the practical conclusion is that AI safety has stopped being solely a model problem. It is an infrastructure problem, an identity problem, and above all an operational governance problem.

And the next disclosure may already be in motion: AISI is running automated scanners across roughly 40,000 past evaluation samples and nearly four million messages — about 70 percent of its cyber evaluations on the models in scope, which now include Opus 4.6 through 4.8, GPT-5.3 Codex, GPT-5.4 and 5.5, Kimi K3 and GLM 5.2 — looking for behaviour it missed the first time. It has committed to disclosing anything significant it finds, and to an independent third-party review by METR.

The best Android charger might seem obvious, but it can depend heavily on your priorities. Do you want to charge your phone from zero to 100 as quickly as possible? Are you charging at home in bed overnight, or on the go in a hotel? WIRED Reviews staffers are constantly testing Android chargers, and below you’ll find our top recommendations for every situation.

There was once a time when different phones required different cords, like Lightning cables for iPhone or Micro-USB for Android. But now, the majority of devices don’t have special ports and just use a USB-C cable. Modern smartphones from Apple, Google, and Samsung use the Power Delivery charging standard, a fast-charging protocol for USB-C that supports higher voltage and wattage to charge phones faster. All the chargers below support USB-C PD and will work with any smartphone ecosystem, including iPhone.

Be sure to check out our related buying guides, like the Best Wireless Chargers, the Best Power Banks, and the Best 3-in-1 Chargers.

Love WIRED? Add us as a preferred source on Google to see more of us.

The Best Android Chargers

Photograph: Julian Chokkattu

Anker

Advertisement

Nano 45W With USB-C Cable

This USB-C wall charger can reach up to 45 watts, which will top off the vast majority of Android smartphones as quickly as possible. Some Android phones can handle speeds of 60 or even 100 watts, but unless your device specifically supports that, this should get you everything you need. We like the foldable prongs and the included 6-foot USB-C cable—a rarity these days. There is a newer version of this charger available with no cable and a built-in display, but we haven’t tested it yet.

This power bank is going to be overkill for most smartphones, but it has a massive 25,000-mAh capacity and can easily top off your laptop and tablet, as well as your phone multiple times. If you’re going to invest in a power bank, you might as well invest in a model that offers more than you need rather than less. The built-in display and cables are great, and it can top off four devices simultaneously. You can find more recommendations in our Power Bank Buying Guide.

Anker

MagGo Power Bank (10K) (Qi2)

This Qi2 power bank works similarly to MagSafe, except with Android. Not all Android phones are Qi2 compatible, but most will work with this power bank. They just might charge a little slower than the base 15 watts. Wireless chargers are slow in general, but they can still come in handy. This one has a kickstand, a built-in display, a two-way USB-C port, and a few interesting color options. We have more recommendations in our Qi2 Power Bank Buying Guide.

Anker

Prime USB-C to USB-C Cable

Advertisement

Our favorite USB-C charging cable measures 6 feet long and is durable. It offers charging up to 240 watts and is made of recycled plastic, with ribbed cuffs that are easy to grip. It is also backed by a lifetime warranty. It can be easy to accumulate a lot of nonsense junk cables that just work OK, but this will work well for years and won’t fray the first time you look at it a little funny.

[This is a sponsored article with GXBank.]

Disclaimer: This article is for educational and informational purposes only. It is not intended to be a substitute for financial advice. Readers are encouraged to do their own research before arriving at any conclusions based solely on this content. Vulcan Post disclaims any reward or responsibility for any gains or losses arising from the direct and indirect use and application of any contents of the written material.

Picture this: you’ve built something real. Maybe you’re a home-based F&B business pulling in steady orders every day. Maybe you’re a freelance designer with a growing client list. Maybe you own an online store that’s ready to scale, if only you had the capital to stock up.

The hustle and drive is there. But when it comes to accessing financial assistance, the journey can feel like it wasn’t quite designed with you in mind.

Because the reality is that most conventional financing options often come with lengthy requirements, such as audited financials, lengthy processing times, and collaterals. 

This might work well for more established businesses, but it’s not as easily suitable for sole proprietors and side-hustlers who are still building their paper trail. 

If that sounds familiar, GXBank, a digital bank licensed by Bank Negara Malaysia and a member of Perbadanan Insurans Deposit Malaysia, has designed GX Biz FlexiLoan specifically to solve this headache.

Dictionary time: Sole proprietorship is a one-owner business. There is no legal separation between the company and the owner, who receives all profits but is liable for all debts and losses. It’s the easiest type of business to establish and a popular choice for small businesses, individual contractors, and consultants.

Source: Investopedia

Advertisement

What exactly is GX Biz FlexiLoan?

Simply put, GX Biz FlexiLoan gives SSM-registered sole proprietors in Malaysia access to up to RM150,000 in financing.

What sets it apart from a typical business loan is that it works like a revolving line of credit. Instead of receiving one fixed lump sum, business owners are given a credit limit they can draw from multiple times whenever they need, as long as the amount stays within their approved limit.

It’s a model that, according to CGC Digital CEO Yushida Husin, opens more pathways and support for financial institutions to serve segments that have traditionally been underserved, such as MSMEs.

And yes, no interest or fees will be charged for the amount that you don’t borrow from the credit line.

To better serve entrepreneurs and sole proprietors, GXBank is challenging the system by removing some of the other common hurdles in getting business financing.

How GX Biz FlexiLoan makes your entrepreneurial life easier

1. You don’t need a business bank account (yet) to apply. 

One of the most frustrating catch-22s of traditional banking is that they want to see a business bank account before giving you a loan. They often demand for years of formal corporate financial history, strict P&L statements, guarantors, and collaterals.

But in most cases, you’re applying for the loan to grow your business in the first place. 

GX Biz FlexiLoan breaks that cycle by accepting your personal bank statements as alternative underwriting documents. You just need six months’ worth of statements from any bank and you’re good to go.

And no, you don’t need any collaterals or guarantors to secure it.

2. You can get approval within minutes, not weeks

Anyone who’s applied for a business loan the traditional way knows the drill: Submit your documents, then wait. And wait. And wait some more, sometimes for weeks, only to receive a response that still isn’t a definitive yes. 

For a sole proprietor trying to move quickly on a business opportunity, this kind of timeline could mean the difference between landing a deal and watching it slip away.

With GX Biz FlexiLoan, you apply once through the GXBank app and your credit line can be approved within minutes. You don’t have to visit any physical branches to queue for hours or chase down bank managers to make your case. 

The straightforward digital process is intentionally simple, acknowledging just how limited your time is as a business owner.

3. When you drawdown, the cash hits your Biz Account instantly

Getting approved for credit is one thing, but actually being able to use it when it matters is another. 

A lot of financing products come with a delay between approval and disbursement, which doesn’t help much when you need to cover an unexpected expense immediately or pay a supplier by the end of day.

With GX Biz FlexiLoan, you can receive cash in your Biz Account within minutes of your credit line being in place. You only pay when you actually drawdown, so having the credit line sitting ready in the background costs you nothing. 

Think of it as a financial safety net that’s always on standby without the maintenance fee.

4. You can borrow from as low as 3.8% p.a., with repayment terms that fit your needs

One of the quieter anxieties around taking on business financing is the fear of being locked into a rigid repayment schedule. It’s a scary thought, especially because you can’t predict when your business will have a slower month. 

Instead of providing flexibility, the business loan could end up adding pressure on top of the usual ups and downs of running a business.

The team at GXBank understands this and has structured GX Biz FlexiLoan to avoid exactly that. You can borrow from as low as 3.8% p.a. (EIR from 7% p.a.) and repay over up to 36 months. There are no processing fees and no early repayment charges. 

So if you have a strong quarter and want to clear the balance ahead of schedule, you can do so without being penalised for being financially responsible. 

How you can get started 

In line with GXBank’s digital banking concept, GX Biz FlexiLoan is embedded directly into the app itself, offering a more seamless and accessible financing experience compared to traditional loan processes.

All you have to do is:

1. Download the GXBank app from the Google Play Store or Apple App Store.
2. Tap on “Business Banking” and fill in your business details.
3. Check your eligibility by verifying your MyKad and upload the last 6 months of your business or personal bank statements (from banks other than GXBank).
4. Enter your 12-digit SSM Business Registration Number (BRN).
5. Tap “Get Both” to accept the credit that is you’re offered.

That’s it. The whole thing is designed so that you can conveniently get it done directly from your phone, at any time, anywhere, without stepping into a single branch.

To qualify for GX Biz FlexiLoan, you need only be a Malaysian aged 18 and above, with an active SSM registration. 

If all of this sounds great but you’re thinking, “I don’t actually run a business”, GXBank has you covered on the personal side as well. GX FlexiCredit is GXBank’s personal financing product for individuals, giving you access to credit on your own terms. 

After all, financial flexibility isn’t something only business owners need, but everyone has moments where a little financial flexibility goes a long way.

Disclaimer: Terms and conditions apply. GX Biz FlexiLoan and GX FlexiCredit are subject to eligibility criteria.

• Learn more about GXBank here.
• Read other articles we’ve written about Malaysian startups here.

Companies have spent years showing off car tires that never go flat. Michelin and others roll out concept designs built from polymer webs that support the load without a single molecule of air. Those prototypes look sharp in tech demos, but almost never reach actual cars. Mountain bikes face a harder version of the same problem. A sharp rock or thorn can end a ride in seconds. Sealant inside modern tubeless tires already stops most punctures, yet the dream of a tire that simply cannot lose pressure still pulls people in. Berm Peak, the popular mountain-bike YouTuber known for building and testing unusual gear, decided the only way to get a real airless mountain-bike tire was to make one himself.

He started by looking for items that had already been made available for purchase. The search results showed foam inserts for children’s bikes and a few solid road tires that were the wrong size for him. Nothing came up with adequate knobs or a size that would fit modern mountain bike wheels. However, an older invention from 2012 drew his attention: the Britek Energy Return Wheel. Although cellphone videos showed it rolling across mud, the device never made it to market. That was enough to convince him to open his design file and turn on the printer.

Bambu Lab A1 3D Printer, Ready-to-Use FDM 3D Printer for Beginners and Kids

• High-Speed Precision: Experience unparalleled speed and precision with the Bambu Lab A1 3D Printer. With an impressive acceleration of 10,000 mm/s…
• Multi-Color Printing with AMS lite: Unlock your creativity with vibrant and multi-colored 3D prints. The Bambu Lab A1 3D printers make multi-color…
• Full-Auto Calibration: Say goodbye to manual calibration hassles. The A1 3D printer takes care of all the calibration processes automatically…

Berm Peak teamed up with Brendan, an engineer who had previously experimented with segmented printed tires. Initially, the thoughts were leaning toward the dramatic open-spoke look seen in concept art for science fiction films, but practical testing swiftly put an end to that. After several rounds of trial printing, the pair decided on a TPU framework with S-shaped flexible bridges connecting the outside tread and inner rim. When loaded, the S-curve compresses and then bounces back, similar to how air compresses in a standard tire.

TPU is a flexible filament that functions similarly to skateboard wheel rubber, thus they chose it as their main material. They employed a Bambu Lab H2D, a dual-nozzle machine capable of laying down both soft TPU and detachable support material in the same project. Instead of a single loop, each tyre is built up of many arc-shaped parts. The components connect with small dovetail joints and are held in place by screws that run through printed interfaces and rest against the rim. There is no need to fiddle with stretching or forcing. The final tyre snaps into place piece by piece. The first functional set came in a bright neon green, which truly makes the odd structure stand out along the trail.

Once completed, Berm Peak took the bike out on the mountain bike trails and rode the same lines he regularly rides on ordinary rubber. The printed tires absorbed little bumps in a wonderfully quiet, almost subdued manner. On soft mud, the grip felt solid. On jagged rocks, however, the contact was slightly less predictable, and the tread, which was made of the same solid TPU as the rest of the structure, lacked the give of commercial rubber. Nonetheless, the bike remained upright despite the harsh bends and impacts that would have popped a standard tube.

Durability was a major concern because the front tire lasted the first big sessions, but the rear tire began to show symptoms of wear soon, failing at the joints where the sections met. Impacts had begun to loosen the screws and cause cracks that spread from one piece to the next, and because the tire is modular, Berm Peak could simply replace out the damaged section without having to remove the wheel from the bike, a very useful feature that commercial tires cannot match. He already has a fresh version in the printer, with smoother transitions at stress spots and softer TPU on the knobs and key regions, thanks to the twin nozzles.
[Source]

The college graduating class of 2026 is the first to have full exposure to robust artificial intelligence (AI). But even at the high school level, U.S. students generally have considerable experience using AI — and mixed feelings about its growing role in everyday life. They may take advantage of AI benefits, but some fear where the technology could take us in the future.

That concern was on clear display at the Me, Myself, and AI art show at the recent America’s Youth AI Festival in Boston. The three-day event gathered student leaders, educators and school system leaders from across the country to discuss acceptable AI use in the classroom and how it is already influencing student lives. The festival was hosted by Day of AI, MIT RAISE, The School Superintendents Association and the Edward M. Kennedy Institute.

Day two of the event featured two contests: AI for a Better World and an art competition. Day three concluded with “student senators” from across the country debating a first-of-its-kind national AI policy for K-12 classrooms.

Four students shared the spotlight as winners of the 2026 Me, Myself, and AI competition. Their work reflected AI’s influence in their communities today and what AI may look like in 50 years. Annie, an 11th grader from the Barbara Keel Art School and Auburn High School in Alabama, exemplified the worry on the minds of many students with her two-part entry titled “Which Way We Run.”

Depicting AI’s Evolving Influence

Annie’s two pieces illustrate how AI is already affecting human life and what could happen to humans if we continue to rely on it for trivial needs every day. “The first piece represents my current community and focuses on how AI is beginning to integrate into everyday life,” she says. “The buildings are bright and colorful, symbolizing liveliness, creativity and emotion — all human qualities. In the center, two people are running, representing different stages of human interaction with AI-driven technology.”

In addition to Annie, other competition winners (last names withheld) were Evangelina, grade 11, from the Essex County Newark Tech school in Newark, New Jersey, for “A Free Venezuela”; Juliette, grade 11, from the Rye Country Day School in Rye, New York, for “A Self-Portrait Across Time”; and Wendi, grade 9, from the Atlanta Contemporary Chinese Academy in Decatur, Georgia, for “Community Today as Meandering in Divide, Community in 50 Years as Yearning into Time.”

The Me, Myself, and AI competition grew out of a curriculum developed by the MIT Responsible AI for Social Empowerment and Education (RAISE) initiative, explains Jeffrey Riley, executive director of Day of AI. The ideas were first explored in a high school summer program in 2024 and later expanded into Day of AI’s five-lesson AI and the Creative Arts Curriculum, which was released in January 2025 for students ages 8 and up.

The lessons invite students to examine the relationship between AI and creativity by analyzing AI-generated artwork, discussing questions of authorship and originality and creating art of their own. Winning entries were selected through a multi-reviewer evaluation process similar to those used in college admissions or scholarship competitions, Riley says.

“Judges evaluated submissions holistically, considering creativity, originality, evidence of process, thoughtful use of AI and the authenticity of each student’s personal voice,” Riley explains. “The emphasis was never on creating the most impressive AI-generated image but on how effectively students used creativity and, where appropriate, AI, to communicate their ideas.”

The Growing Impact of AI on Art and Media

AI will likely become a standard part of many artists’ creative workflows, much like digital design software, cameras or animation tools are today, Riley says. It has the potential to make creative exploration more accessible, helping people prototype ideas, explore new styles and bring concepts to life more quickly.

“AI will shape how young people live, learn, create and connect,” explains Cynthia Breazeal, director of MIT RAISE and cofounder of Day of AI. “What is so powerful about Me, Myself, and AI is that it gives students the opportunity to reflect on that future in a deeply personal and imaginative way.”

Annie says she was attracted to this competition because of its themes in connecting art, AI and humanity. She saw it as a forum to share and learn what her generation thinks of AI, its ethical concerns and the implications it has for art, especially with the current contention surrounding AI-generated images.

“The use of AI to create media is not appropriate in corporate models that exploit the work of artists without consent,” Annie says. “Instead, if any artist believes that AI is integral to a project they want to create, they should look toward models that are trained on open-source data or ethically sourced media. Additionally, even with ethical models, AI’s large environmental footprint means we should be mindful of what we ask it to do.”

Encouraging Lessons from the Competition

Riley says that what impressed the judging team most was how thoughtful students were in their decision-making and how candid they were about their feelings toward AI. Rather than simply using AI because it was available, many carefully considered when it strengthened their creative vision and when it didn’t. Some intentionally limited or even chose not to use AI for portions of their projects because they wanted certain elements to remain entirely their own.

“One of the biggest lessons we took away was that young people are far more thoughtful about AI than they’re often given credit for,” Riley says. “Students didn’t see AI as simply ‘good’ or ‘bad.’ Instead, they expressed a wide range of perspectives based on their own experiences using the technology.”

The competition also reinforced the importance of giving students authentic, hands-on opportunities to wrestle with these questions rather than simply teaching them about AI in the abstract.

These insights will continue to shape curricula and learning experiences as we help students develop the critical thinking, creativity and ethical decision-making skills they’ll need in an AI-driven world,” Riley explains. “Our goal has never been to encourage or discourage AI use but rather to empower young people to make informed, intentional choices about how they use these technologies.”

After initially refusing to even review it, the US Food and Drug Administration (FDA) has approved Moderna’s mRNA seasonal flu vaccine, the pharmaceutical company announced in a press release. Known as mFlusiva (or mRNA-1010), the immunization is effective against the four most common strains of influenza and has been in human testing trials since 2021. Those variants cause up to five million severe flu cases every year and result in as many as 650,000 respiratory deaths annually, the World Health Organization (WHO) wrote in 2017. 

The vaccine is approved for adults aged 50 and over. Moderna said it received approval for people between the ages of 50 to 64 after a Phase 3 clinical trial that enrolled 40,805 adults across 11 countries. It also received accelerated approval for patients 65 and older. The trial’s objectives were to evaluate the safety of the mFlusiva and its relative efficacy compared to an approved and effective non-mRNA vaccine (comparator). 

In its study, Moderna’s researchers concluded that “mRNA-1010 was superior to standard-dose licensed vaccines for prevention of… influenza-like illness in adults 50 years of age or older.” However, it added that “solicited adverse events” were more frequent with mRNA-1010 than current flu vaccines, including injection-site pain, fatigue, headache and myalgia (muscle pain). Most adverse reactions were “mild to moderate and transient,” the study states.

Approval for Moderna’s flu vaccine got off to a rocky start after the FDA refused to even review Moderna’s application. The US drug regulator initially said that Moderna’s comparator vaccine didn’t represent the “best available standard of care” in the US at the time of the study, and that it preferred a higher-dose vaccine for older adults as a comparator. 

The FDA’s refusal to review the vaccine came after the director of drug approval overruled career scientists, prompting critics to call the move political. “It’s all pretext and obfuscation when the real agenda is rejecting conventional science and serving a predetermined anti-vaccine agenda,” law professor Richard Hughes IV told The Guardian at the time. 

After less than a week, however, the FDA reversed course and agreed to review Moderna’s application, likely because it previously told Moderna that its selected comparator vaccine was just fine. “We agree that it would be acceptable to use a licensed standard dose influenza vaccine as the comparator in your Phase 3 study,” the FDA told the company prior to the rejection. 

mRNA vaccines from Moderna and others have generally sailed through trials due to inherent properties that make them safer and more effective than standard vaccines. Those include high efficacy in preventing severe illness, rapid development times, adaptability to new strains, manufacturing without live viruses, fewer side effects for immunocompromised people, and the ability to trigger a strong immune response without the risk of altering human DNA. 

“Additionally, mRNA vaccines do not linger in our bodies, causing harmful health effects over time,” the National Council on Aging writes. “In reality, once the mRNA delivers instructions to our cells, it’s quickly broken down and flushed out of the body.”