Quote for the day:
“The greatest leader is not necessarily the one who does the greatest things. He is the one that gets the people to do the greatest things.” -- Ronald Reagan
Neuromorphic computing and the future of edge AI
While QC captures the mainstream headlines, neuromorphic computing has
positioned itself as a force in the next era of AI. While conventional AI
relies heavily on GPU/TPU-based architectures, neuromorphic systems mimic the
parallel and event-driven nature of the human brain. ... Neuromorphic hardware
has shown promise in edge environments where power efficiency, latency and
adaptability matter most. From wearable medical devices to battlefield
robotics, systems that can “think locally” without requiring constant cloud
connectivity offer clear advantages. ... As neuromorphic computing matures,
ethical and sustainability considerations will shape adoption as much as raw
performance. Spiking neural networks’ efficiency reduces carbon footprints by
cutting energy demands compared to GPUs, aligning with global decarbonization
targets. At the same time, ensuring that neuromorphic models are transparent,
bias‑aware and auditable is critical for applications in healthcare, defense
and finance. Calls for AI governance frameworks now explicitly include
neuromorphic AI, reflecting its potential role in high‑stakes decision‑making.
Embedding sustainability and ethics into the neuromorphic roadmap will ensure
that efficiency gains do not come at the cost of fairness or
accountability.10 security leadership career-killers — and how to avoid them
“Security has evolved from being the end goal to being a business-enabling
function,” says James Carder, CISO at software maker Benevity. “That means
security strategies, communications, planning, and execution need to be
aligned with business outcomes. If security efforts aren’t returning
meaningful ROI, CISOs are likely doing something wrong. Security should not
operate as a cost center, and if we act or report like one, we’re failing in
our roles.” ... CISOs generally know that the security function can’t be the
“department of no.” But some don’t quite get to a “yes,” either, which means
they’re still failing their organizations in a way that could stymie their
careers, says Aimee Cardwell, CISO in residence at tech company Transcend and
former CISO of UnitedHealth Group. ... CISOs who are too rigid with the
rules do a disservice to their organizations and their professional prospects,
says Cardwell. Such a situation recently came up in her organization, where
one of her team members initially declined to permit a third-party application
from being used by workers, pointing to a security policy barring such apps.
... CISOs who don’t have a firm grasp on all that they must secure won’t
succeed in their roles. “If they don’t have visibility, if they can’t talk
about the effectiveness of the controls, then they won’t have credibility and
the confidence in them among leadership will erode,” Knisley says.A CIO's Evolving Role in the Generative AI Era
The dual mandate facing CIOs today is demanding but unavoidable. They must
deliver quick AI pilots that boards can take to the shareholders while also
enforcing guardrails on security, ethics and cost aspects. Too much caution
can make CIOs irrelevant. This balancing act requires not only technical
fluency but also narrative skill. The ability to translate AI experiments into
business outcomes that CEOs and boards can trust can make CIOs a force. The
MIT report highlights another critical decision point: whether to build or
buy. Many enterprises attempt internal builds, but externally built AI
partnerships succeed twice as often. CIOs, pressured for fast results, must be
pragmatic about when to build and when to partner. Gen AI does not - and never
will - replace the CIO role. But it demands corrections. The CIO who once
focused on alignment must now lead business transformation. Those who succeed
will act less as CIOs and more as AI diplomats, bridging hype with pragmatism,
connecting technological opportunities to shareholder value and balancing the
boardroom's urgency with the operational reality. As AI advances, so does the
CIO's role - but only if they evolve. Their reporting line to the CEO
symbolizes greater trust and higher stakes. Unlike previous technology
cycles, AI has brought the CIO to the forefront of transformation. Building an AI Team May Mean Hiring Where the Talent Is, Not Where Your Bank Is
Much of the adaptation of banking to AI approaches requires close collaboration
between AI talent with people who understand how the banking processes involved
need to work. This will put people closer together, literally, to facilitate
both quick and in-depth but always frequent interactions to make collaboration
work — paradoxically, increased automation needs more face-to-face dealings at
the formative stages. However, the "where" of the space will also hinge on where
AI and innovation talent can be recruited, where that talent is being bred and
wants to work, and the types of offices that talent will be attracted to. ...
"Banks are also recruiting for emerging specialties in responsible AI and AI
governance, ensuring that their AI initiatives are ethical, compliant and
risk-managed," the report says. "As ‘agentic AI’ — autonomous AI agents — and
generative AI gain traction, firms will need experts in these cutting-edge
fields too." ... Decisions don’t stop at the border anymore. Jesrani says that
savvy banks look for pockets of talent as well. ... "Banks are contemplating
their global strategies because emerging markets can provide them with talent
and capabilities that they may not be able to obtain in the U.S.," says Haglund.
"Or there may be things happening in those markets that they need to be a part
of in order to advance their core business capabilities."
How Data Immaturity is Preventing Advanced AI
Data immaturity, in the context of AI, refers to an organisation’s
underdeveloped or inadequate data practices, which limit its ability to leverage
AI effectively. It encompasses issues with data quality, accessibility,
governance, and infrastructure. Critical signs of data immaturity include
inconsistent, incomplete, or outdated data leading to unreliable AI outcomes;
data silos across departments hindering access and comprehensive analysis, as
well as weak data governance caused by a lack of policies on data ownership,
compliance and security, which introduces risks and restricts AI usage. ... Data
immaturity also leads to a lack of trust in analysis and predictability of
execution. That puts a damper on any plans to leverage AI in a more autonomous
manner—whether for business or operational process automation. A recent study by
Kearney found that organisations globally are expecting to increase data and
analytics budgets by 22% in the next three years as AI adoption scales.
Fragmented data limits the predictive accuracy and reliability of AI, which are
crucial for autonomous functions where decisions are made without human
intervention. As a result, organisations must get their data houses in order
before they will be able to truly take advantage of AI’s potential to optimise
workflows and free up valuable time for humans to focus on strategy and design,
tasks for which most AI is not yet well suited.
From Reactive Tools to Intelligent Agents: Fulcrum Digital’s AI-First Transformation
To mature, LLM is just one layer. Then you require the integration layer, how you integrate it. Every customer has multiple assets in their business which have to connect with LLM layers. Every business has so many existing applications and new applications; businesses are also buying some new AI agents from the market. How do you bring new AI agents, existing old systems, and new modern systems of the business together — integrating with LLM? That is one aspect. The second aspect is every business has its own data. So LLM has to train on those datasets. Copilot and OpenAI are trained on zillions of data, but that is LLM. Industry wants SLM—small language models, private language models, and industry-orientated language models. So LLMs have to be fine-tuned according to the industry and also fine-tuned according to their data. Nowadays people come to realise that LLMs will never give you 100 per cent accurate solutions, no matter which LLM you choose. That is the phenomenon customers and everybody are now learning. The difference between us and others: many players who are new to the game deliver results with LLMs at 70–75 per cent. Because we have matured this game with multiple LLMs coexisting, and with those LLMs together maturing our Ryze platform, we are able to deliver more than 93–95 per cent accuracy.You Didn't Get Phished — You Onboarded the Attacker
Many organizations respond by overcorrecting: "I want my entire company to be
as locked down as my most sensitive resource." It seems sensible—until the
work slows to a crawl. Without nuanced controls that allow your security
policies to distinguish between legitimate workflows and unnecessary exposure,
simply applying rigid controls that lock everything down across the
organization will grind productivity to a halt. Employees need access to do
their jobs. If security policies are too restrictive, employees are either
going to find workarounds or continually ask for exceptions. Over time, risk
creeps in as exceptions become the norm. This collection of internal
exceptions slowly pushes you back towards "the castle and moat" approach. The
walls are fortified from the outside, but open on the inside. And giving
employees the key to unlock everything inside so they can do their jobs means
you are giving one to Jordan, too. ... A practical way to begin is by piloting
ZSP on your most sensitive system for two weeks. Measure how access requests,
approvals, and audits flow in practice. Quick wins here can build momentum for
wider adoption, and prove that security and productivity don't have to be at
odds. ... When work demands more, employees can receive it on request through
time-bound, auditable workflows. Just enough access is granted just in time,
then removed. By taking steps to operationalize zero standing privileges, you
empower legitimate users to move quickly—without leaving persistent privileges
lying around for Jordan to find.
OT Security: When Shutting Down Is Not an Option
some of the most urgent and disruptive threats today are unfolding far from the keyboard, in operational technology environments that keep factories running, energy flowing and transportation systems moving. In these sectors, digital attacks can lead to physical consequences, and defending OT environments demands specialized skills. Real-world incidents across manufacturing and critical infrastructure show how quickly operations can be disrupted when OT systems are not adequately protected. Just this week, Jaguar Land Rover disclosed that a cyberattack "severely disrupted" its automotive manufacturing operations. ... OT environments present challenges that differ sharply from traditional IT. While security is improving, OT security teams must protect legacy control systems running outdated firmware, making them difficult to patch. Operators need to prioritize uptime and safety over system changes; and IT and OT teams frequently work in silos. These conditions mean that breaches can have physical as well as digital consequences, from halting production to endangering lives. Training tailored to OT is essential to secure critical systems while maintaining operational continuity. ... An OT cybersecurity learning ecosystem is not a one-time checklist but a continuous program. The following elements help organizations choose training that meets current needs while building capacity for ongoing improvement.Connected cars are racing ahead, but security is stuck in neutral
Connected cars are essentially digital platforms with multiple entry points
for attackers. The research highlights several areas of concern. Remote access
attacks can target telematics systems, wireless interfaces, or mobile apps
linked to the car. Data leaks are another major issue because connected cars
collect sensitive information, including location history and driving
behavior, which is often stored in the cloud. Sensors present their own set of
risks. Cameras, radar, lidar, and GPS can be manipulated, creating confusion
for driver assistance systems. Once inside a vehicle, attackers can move
deeper by exploiting the CAN bus, which connects key systems such as brakes,
steering, and acceleration. ... Most drivers want information about what data
is collected and where it goes, yet very few said they have received that
information. Brand perception also plays a role. Many participants prefer
European or Japanese brands, while some expressed distrust toward vehicles
from certain countries, citing political concerns, safety issues, or perceived
quality gaps. ... Manufacturers are pushing out new software-defined features,
integrating apps, and rolling out over the air updates. This speed increases
the number of attack paths and makes it harder for security practices and
rules to keep up.
