Troubleshooting Common Issues

Are you solving the right problem? Now that probability and uncertainty is creeping into previously deterministic systems, it's time to talk about errors -- those bad conclusions you're about to jump to. Everyone in data science knows about Type I and Type II errors: 1️⃣ Type I Error = False positive. You thought you found something actionable, but it was noise. 2️⃣ Type II Error = False negative. You missed a real signal and failed to change course. But the one that should really keep you up at night is the Type III Error: ✔️ All the right math, beautiful dashboards, flawless execution… ❌ Solving the wrong problem. 3️⃣ Type III Error = Wrong positive. It's... The boardroom high-five that shouldn’t have happened. The KPI that looks impressive, but delivers no actual value. Organizations love to ask: “What does the data say?” But often they're skipping the more important question: “Are we asking the right question?” The most dangerous AI/ML system isn’t the one that breaks. It’s the one that works perfectly—on a goal that shouldn't exist in the first place. That’s why I keep saying: “Skilled decision-making is a must-have for effective AI and data science.” Decision intelligence is how you elevate the judgment and framing skills required to turn information into better action. And that’s where most organizations are weakest. They hire technical folk before the leaders have done their homework and properly clarified the decisions worth making. And the more your systems scale, the more dangerous this becomes. Want to reduce Type III errors? Here’s what that takes: ✅ Start with the decision/action/vision, not the data. ✅ Define what “better” means before you look for insights. ✅ Think through the alternatives before automating anything. ✅ Bring in decision scientists—don’t expect everyone to be one without training. ✅ Watch out for technically flawless projects that deliver suspiciously little impact. Data-driven decisions aren’t the same as data-decorated decisions. Your turn: Have you ever seen a Type III error in the wild? What helped you catch it? If you found this useful, a repost ♻️ makes my heart happy. And a subscription to my newsletter makes my day. decision.substack.com #DecisionIntelligence #DataScience #Leadership #AI #DecisionMaking *Footnote for my fellow statisticians in the room: We statisticians shudder unless the meaning is exactly right, so here's the more proper set of definitions: Type I Error: Incorrectly rejecting the null hypothesis. Leaving a good default action. Type II Error: Incorrectly failing to reject the null hypothesis. Staying with a bad default action. Type III Error: Correctly rejecting the wrong null hypothesis. Wasting your life. If you read this far and were cheered by that footnote, you're the best kind of nerd -- definitely repost ♻️ keep the good stuff alive. Join my newsletter where sensible leaders go for AI and decision science: decision.substack.com

Many are asking me... Should I continue to track "Open Rates" on Cold Emails? It's still no. My answer hasn't changed. I had predicted this about 9 months ago if you want to look back. Why? Analyze the image in the post. Does the position of the "Report as Spam" increase the amount of people who click it by 3 on 1,000 recipients? If you said yes, you agree with me. This is a subtle way Google is asking you for more feedback on the quality of your outbound campaigns. Here are 5 reasons NOT to use Open Tracking for Cold Email: Reason 1: Limits Your Use Of Plain Text Emails Plain Text Emails get superior deliverability. Open Trackers can't be used in Plain Text emails. Reason 2: Inconsistent Tracking Open Trackers identify "opens" differently and ultimately can't prove someone opened the email. Every sequencer has a different way of tracking it. Reason 3: Email Fingerprints Open Trackers provide a fingerprint for your domain reputation. It's shared amongst everyone using the sequencer your company uses. Do you want to be part of this group? Reason 3: Misleading Data Secure Email Gateways open emails for their users to protect their privacy. Budget has increased significantly here and will continue to go up. Most of these systems will put your email in spam because of it. Reason 4: Easy To Block Even simple rules can block emails with open trackers. No AI required. It's simple. Reason 5: Bad Metric Teams and internet gurus are obsessed with open tracking. However, it doesn't mean your email has been opened. It could mean that, but it depends who you emailed. Here are 3 Insider Tips to Improve Deliverability Today: Insider Tip #1: Send to less technical audiences. This isn't my favorite advice to give. However, less technical audiences hit the report as spam button less. Insider Tip #2: Send to companies without Proofpoint, Cisco, and Mimecast MX Records. Prioritize companies invested in email security systems lower than ones who don't. Use LeadMagic to figure out what the company uses in the email finder. Insider Tip #3: Use LeadMagic's New Features on MX Detection & Valid_Catch_All Status to prioritize who to send to first. Prioritize valid (mail server checked) > catch_all. Use valid_catch_all status from LeadMagic which detects if the email has been found other ways. Prioritize Google or Microsoft email servers higher than Proofpoint, Cisco, and Mimecast email servers. This will lead to better delivery & reply rates. p.s. open tracking is not dead for email marketing, but that's not what I am talking about.

As an MLOps platform, we started by helping organizations implement responsible AI governance for traditional machine learning models. With principles of transparency, accountability, and oversight, our Guardrails enabled smooth model development. However, governing large language models (LLMs) like ChatGPT requires a fundamentally different approach. LLMs aren't narrow systems designed for specific tasks - they can generate nuanced text on virtually any topic imaginable. This presents a whole new set of challenges for governance. Here are some key components for evolving AI governance frameworks to effectively oversee large language models (LLMs): 1️⃣ Usage-Focused Governance: Focus governance efforts on real-world LLM usage - the workflows, inputs and outputs - rather than just the technical architecture. Continuously assess risks posed by different use cases. 2️⃣ Dynamic Risk Assessment: Identify unique risks presented by LLMs such as bias amplification and develop flexible frameworks to proactively address emerging issues. 3️⃣ Customized Integrations: Invest in tailored solutions to integrate complex LLMs with existing systems in alignment with governance goals. 4️⃣ Advanced Monitoring: Utilize state-of-the-art tools to monitor LLMs in real-time across metrics like outputs, bias indicators, misuse prevention, and more. 5️⃣ Continuous Accuracy Tracking: Implement ongoing processes to detect subtle accuracy drifts or inconsistencies in LLM outputs before they escalate. 6️⃣ Agile Oversight: Adopt agile, iterative governance processes to manage frequent LLM updates and retraining in line with the rapid evolution of models. 7️⃣ Enhanced Transparency: Incorporate methodologies to audit LLMs, trace outputs back to training data/prompts and pinpoint root causes of issues to enhance accountability. In conclusion, while the rise of LLMs has disrupted traditional governance models, we at Katonic AI are working hard to understand the nuances of LLM-centric governance and aim to provide effective solutions to assist organizations in harnessing the power of LLMs responsibly and efficiently. #LLMGovernance #ResponsibleLLMs #LLMrisks #LLMethics #LLMpolicy #LLMregulation #LLMbias #LLMtransparency #LLMaccountability

Toyota's competitors thought Taiichi Ohno was insane. He was letting line workers stop production over a scratched fender. They stopped laughing when Toyota dominated the industry for the next 50 years. Here's what happened: After World War II, Ohno visited auto factories in Detroit. He expected to learn from the best. Instead, he found rework areas filled with broken pieces and leftover parts. Waste everywhere, hiding in plain sight. He went back to Toyota City and did something nobody had tried before. At every station on the assembly line, he hung a rope called the Andon cord. The instruction was simple: If you see a defect pull the cord. The line slows or stops. Engineers, workers, and suppliers huddle up and fix it on the spot. Detroit thought he was crazy. "How can you build thousands of cars a day if you stop the line for every little scratch?" Ohno's answer was simple: A scratched fender is an early warning that a piece of equipment is failing. Fix the scratch today, and you prevent 500 defective fenders next week. Ignore it, and you're building rework areas just like Detroit. That's the difference between firefighting and fire prevention. Most operations leaders I talk to are stuck in firefighting mode. They walk past small defects every day because the line has to keep moving. But those small defects are talking to you. They're telling you exactly where your next big quality failure will come from. How many "minor" defects is your team walking past today that are actually early warning signs?

LLM hallucinations aren't bugs, they're compression artefacts. And we just figured out how to predict them before they happen. 400 stars in one week, the reception has been unreal. Our toolkit is open source and anyone can use it. https://lnkd.in/e4s3X8GK When your LLM confidently states that "Napoleon won the Battle of Waterloo," it's not broken. It's doing exactly what it was trained to do: compress the entire internet into model weights, then decompress on demand. Sometimes, there isn't enough information to perfectly reconstruct rare facts, so it fills gaps with statistically plausible but wrong content. Think of it like a ZIP file corrupted during compression. The decompression algorithm still runs, but outputs garbage where data was lost. The breakthrough: We proved hallucinations occur when information budgets fall below mathematical thresholds. Using our Expectation-level Decompression Law (EDFL), we can calculate exactly how many bits of information are needed to prevent any specific hallucination, before generation even starts. This resolves a fundamental paradox: LLMs achieve near-perfect Bayesian performance on average, yet systematically fail on specific inputs. We proved they're "Bayesian in expectation, not in realisation", optimising average-case compression rather than worst-case reliability. Why this changes everything? Instead of treating hallucinations as inevitable, we can now: Calculate risk scores before generating any text Set guaranteed error bounds (e.g. 95%) Know precisely when to gather more context vs. abstain The full preprint is being released on arXiv this week. Until then, read the preprint PDF we uploaded here: https://lnkd.in/eRf_ecu3 The toolkit works with any OpenAI-compatible API. Zero retraining required. Provides mathematical SLA guarantees for compliance. Perfect for healthcare, finance, legal, anywhere errors aren't acceptable. The era of "trust me, bro" AI is ending. Welcome to bounded, predictable AI reliability. Big thanks to Ahmed K. Maggie C. for all the help putting this + the repo together! #AI #MachineLearning #ResponsibleAI #OpenSource #LLM #Innovation

Understanding Losses in Solar Plants and Types of Solar Plant Losses, why it is important ? Solar power plants are designed to maximize energy production, but various losses can reduce their efficiency and overall energy yield. Understanding these losses is crucial for improving the performance, reliability, and financial viability of solar energy projects Solar plant losses can be categorized into the following types: 1. Irradiance Losses Shading Losses: Obstructions like buildings, trees, or other solar panels can block sunlight, reducing energy output. Soiling Losses: Accumulation of dirt, dust, or bird droppings on panels reduces the amount of sunlight reaching the solar cells. Atmospheric Losses: Variations in atmospheric conditions like clouds or haze can scatter or absorb sunlight, reducing irradiance. 2. Module-Level Losses Mismatch Losses: Differences in the performance of individual solar cells or modules (due to manufacturing variations or shading) lead to energy losses. Temperature Losses: High temperatures reduce the efficiency of photovoltaic (PV) cells, as their performance decreases with heat. Degradation Losses: Over time, solar panels degrade, producing less energy compared to their initial performance. 3. Inverter Losses Conversion Losses: Inverters convert DC power from solar panels to AC power for grid usage. Inefficiencies in this conversion process cause energy losses. Inverter Downtime: Malfunctions or maintenance-related downtime in inverters can lead to energy production losses. 4. Wiring and Electrical Losses Ohmic Losses: Resistance in electrical wiring causes a portion of the energy to dissipate as heat. Connection Losses: Poor-quality or loose electrical connections can lead to energy losses. Transformer Losses: Transformers used to step up or step down voltage introduce inefficiencies. 5. Operational Losses Maintenance Issues: Delayed or inadequate maintenance can lead to prolonged periods of reduced energy production. Monitoring Gaps: Without real-time monitoring, underperforming components may go unnoticed. 6. Environmental and External Factors Weather Variability: Seasonal and daily variations in sunlight availability affect overall energy production. Grid Curtailment: At times, grid operators may restrict the injection of power from solar plants, leading to energy losses. *Why Understanding Solar Plant Losses Is Important* 1. Maximizing Efficiency By identifying and addressing losses, operators can enhance the overall efficiency of the solar plant, ensuring optimal energy production. Improving Financial Returns 2. Reducing losses directly translates to higher energy output, improving revenue generation and return on investment. 3. Long-Term Reliability Regular monitoring and mitigation of losses ensure that solar plants operate reliably over their intended lifespan. 4. Environmental Impact Improved energy yield means more clean energy is produced, reducing dependence on fossil fuels.

PSA 1: Gmail did not kill Email Open Tracking. PSA 2: Email Open Tracking is dead for years. Let’s unpack this. Recently, a screenshot of Gmail blocking images has been circulating on LinkedIn, accompanied by alarmist claims that this spells the end for open tracking. But here’s the truth: Yes, email open tracking relies on images being loaded — typically, by detecting whether a tracking pixel (a tiny, transparent 1x1 pixel unique to each email recipient) has been downloaded. No, Gmail did not just start blocking these pixels. The screenshot actually shows a specific scenario: Gmail blocks images when it identifies an email as likely spam or a scam. Google does this to protect you from being tracked by malicious senders, and it’s been working this way for years. So, does this mean your open tracking is safe and sound? Not really. While Gmail hasn’t started blocking all your tracking pixels, other Email Service Providers already do. Open tracking is frequently blocked by B2B email server admins, often inaccurate due to security bots, and impacted by privacy settings and browser extensions. So, is Email Open Tracking useless? Well… maybe. If you’re using it as a high-level trend marker for opens, it might still offer some value. But if you’re relying on it for behavioral decision-making or key performance indicators (KPIs), especially in the B2B market, it’s largely ineffective. What should you do instead? Click tracking is a better option — although still not perfect, especially due to security bots in the B2B market. Ultimately, the best approach is to focus on the final goal of your email. Why are you sending it? If it’s to sell a product, track product purchases instead. More to come, so keep on analysing #MarketingCloud #SalesforceOhana and #MarketingChampions!

OKRs are widespread but often implemented wrong. Here's 5 first principles when implementing OKRs: 1️⃣ OKRs should align, not cascade The most prolific mistake I see when implementing OKRs is making them cascade in a top-down fashion. This not only remove autonomy, it reduces flexibility as KRs get inherited meaning that pivoting or adapting requires moving back 'up the chain' and to change higher level OKRs. Instead you should think of aligning your OKRs like a constellation of stars, not a linear ladder. 2️⃣ Outcomes over outputs OKRs should describe outcomes not outputs. Avoid having KRs or Objectives that says something like; “Launch app X” 3️⃣ OKRs are a model of you Strategy OKRs don't exist in a vacuum. They should represent the strategic levers you are choosing to pull in order to forward and measure progress towards your strategy. Therefore you must have a strategy first before you set your OKRs. 4️⃣ Customer AND Business impact A common mistake with OKRs are they can become too internally (business) focused. I see companies describe OKRs which state things like "Grow DAU by x% YoY" and "Reach $20M in ARR" etc. And whilst all appropriate goals to strive towards we can easily become too internally focused on forwarding our own agenda and neglect the customer. Therefore the OKRs should balance our goals as a business as well as outcomes for the customer. If we do want to reach a certain ARR then we should be asking ourselves - what customer outcomes might help us get there? These must also become part of our OKRs. 5️⃣ OKRs ≠ KPIs You need both. OKRs = the things we want to change. KPIs = measures of health KPIs are key indicators of performance, they represent the measures that would indicate a performing and healthy product or business. For example, we may want to monitor 'churn', 'conversion' our CAC and costs base, revenue, etc. However we may only want to reduce 'churn' for example. This would then become your OKR. Whilst the other metrics are still important and therefore remain as your KPIs. ===== More on OKRs: 📺 Best Practices on Implementing OKRs (https://lnkd.in/gW6Nms_V) 🗒️ Latest newsletter issue: OKRs vs KPIs (https://lnkd.in/gbmTh_hf)

BOAT auditor, BSR, made some strong observations in the CARO report (mentioned in DRHP filed) , and the points are worth everyone’s attention. 👉 Mismatch in financial statements submitted to banks & financial institutions vs. the company’s own books. (A simple rule: what you show outside should match what you maintain inside.) 👉 Short-term funds used for long-term purposes across multiple subsidiaries. This is like using your monthly budget for a yearly commitment — risky and unsustainable. 👉 Potential going-concern issues — the auditor believes the company may struggle to meet certain liabilities in a subsidiary. 👉 Audit trail deficiencies — signalling weaknesses in internal systems and controls. 👉 Director remuneration exceeding limits under the Companies Act — another governance red flag. These observations may sound harsh, but bold auditing is necessary. It helps companies build stronger controls. While working with Big 4 and Big 6, I had seen similar issues in many startups that I had audited and found the same kinds of gaps. Transparent reporting doesn’t damage a company; it strengthens it. Issues come when problems stay hidden, not when they are highlighted. If startups want investor confidence, strong financial discipline and robust controls are not optional — they are essential. PS : Attaching screenshot from DRHP filed by Boat. #boat #audit #cafinal #castudents #caaspirants #icaistudents