Real problems worth solving

The CCDBG subsidy cliff means a family earning just above 85% of state median income loses their entire childcare subsidy with no taper -- a raise of $2,000/year in wages can trigger a loss of $10,000+ in annual childcare assistance. Parents rationally turn down promotions and raises to stay below the threshold, trapping them in lower-earning positions indefinitely. This cliff exists because federal law caps eligibility at 85% of state median income with no mandatory phase-out schedule, and most states set their cutoffs even lower. The result is a poverty trap where the families most motivated to climb out of low-income brackets are penalized the hardest for doing so.

Dealers affix a second window sticker listing pre-installed add-ons like nitrogen tire fills ($99), paint protection film ($399), VIN etching ($199), and fabric sealant ($299) that were never requested by the buyer and cost the dealer under $50 total. These items inflate the price $800-$1,400 above MSRP and are presented as non-negotiable because they are 'already installed.' This persists because there is no federal law prohibiting dealers from pre-installing accessories and marking them up 10x, the add-ons are physically on the car so the dealer claims they cannot be removed, and consumers arriving at the lot after an online price quote feel pressured to accept rather than walk away after investing hours in the visit.

The dependent care FSA limit remained at $5,000 per household for nearly 40 years, from 1986 through 2025, while average childcare costs rose from roughly $3,000/year to over $14,000/year. That $5,000 cap in 1986 dollars would be $13,700 today, meaning families lost roughly $8,700 in pre-tax savings capacity every year due to legislative inaction. Even the 2025 increase to $7,500 still covers barely half of average infant care costs. This persists because the FSA cap requires an act of Congress to change, and childcare policy consistently loses priority to headline tax issues, so families silently absorbed the erosion for four decades.

Dealerships charge a 'documentation fee' for processing the sale paperwork that ranges from $85 in California (which caps it) to over $1,000 in Florida and other uncapped states, despite the actual cost of printing and filing being under $50. The fee is pure margin disguised as a cost-of-doing-business line item, and because it is applied after price negotiation, consumers who fought hard on the vehicle price discover the savings were clawed back at the finance desk. This persists because most states have no cap on doc fees, dealers are only required to charge the same fee to all customers (not a reasonable fee), and the fee is buried in a stack of closing documents where it is easy to miss.

State regulations mandate one caregiver for every 3 to 4 infants, meaning a single infant room of 8 babies requires 2-3 full-time staff whose wages alone exceed what most families can afford. This ratio is non-negotiable for safety, but it makes infant care the most expensive age bracket -- averaging $14,400/year nationally and over $28,000 in states like Massachusetts. The structural reason this persists is that unlike K-12 education, infant care receives almost no public subsidy to offset the labor intensity, so 100% of the staffing cost falls on parents. The result is 12-18 month waitlists in major metros because providers cannot profitably add infant rooms without charging rates that price out most families.

Credit bureau dispute investigations are handled by a skeleton crew of low-wage workers, many based overseas, who process disputes at industrial volume with no meaningful review. TransUnion, for example, had just 171 workers responsible for investigating disputes across 38 million line items in 2021. Each worker processes disputes in minutes using scripted procedures, verifying the furnisher's data against the furnisher's own records rather than evaluating the consumer's evidence. Consumers who write detailed dispute letters with bank statements and court orders attached get the same assembly-line treatment as frivolous disputes. The result is that legitimate errors survive investigation at rates far higher than they should. This persists because bureaus are paid by data furnishers, not consumers, so investing in thorough investigations is a cost center that reduces profit margins, and the per-dispute penalties under the FCRA are too small to change the calculus.

Dealers deliberately undervalue trade-ins by $2,000-$5,000 below fair market value, using the four-square negotiation method to bundle the trade-in, purchase price, down payment, and monthly payment into a single opaque negotiation. This matters because the average consumer cannot mentally arbitrage four variables simultaneously, so the dealer shifts margin from the visible purchase discount to the invisible trade-in deficit. The practice persists structurally because there is no legal requirement to offer fair market value on a trade, the appraisal process is entirely subjective with no standardized methodology, and most buyers lack the time or knowledge to get competing offers from CarMax or Carvana before sitting at the dealer's desk.

Customers can report an order as 'never delivered' through the app and receive an instant refund with no investigation — the driver has no way to dispute it beyond a grainy GPS screenshot, and the restaurant eats the food cost. Organized rings exploit this by placing high-value orders, claiming non-delivery, and reselling the food, costing individual restaurants thousands per month. Drivers accumulate 'non-delivery' strikes on their accounts for orders they provably completed, and enough strikes lead to deactivation with no meaningful appeals process. Platforms default to instant refunds because fighting chargebacks costs more than absorbing them, and the fraud cost is distributed across restaurants and drivers rather than the platform itself.

Rent is the single largest monthly expense for 44 million American renter households, yet only 27% of property managers report payment data to credit bureaus. A renter paying $2,000/month on time for a decade builds no credit history from that payment, while a borrower with a $200 secured credit card builds a robust file. Tenants who want their rent reported must either convince their landlord to sign up with a reporting service (most refuse because it costs money and adds administrative burden) or pay a third-party rent-reporting service $5-10/month out of pocket to relay their own payment data. Even then, most services only report to one or two bureaus. This asymmetry persists because rent reporting creates no revenue for bureaus comparable to what lenders pay for credit data, small landlords lack the infrastructure to integrate with reporting systems, and no federal law mandates rent reporting.

Dealership service advisors routinely condense manufacturer maintenance schedules by 3-5x, recommending fluid flushes, brake pad replacements, and alignments far earlier than the owner's manual specifies, using misleading language like 'factory required maintenance.' A transmission flush recommended at 12,000-mile intervals instead of the manual's 60,000-mile interval generates four unnecessary services at ~$200 each, costing the owner $800 in pure waste. This persists because service advisors work on commission or are evaluated on per-ticket revenue, fixed operations now account for 39.6% of total dealership gross profit at 45-55% margins, and most consumers never cross-reference the dealer's recommendation against their owner's manual.

Delivery drivers idle their cars for 3-5 hours per shift while waiting for orders in restaurant parking lots, because turning off the engine means losing air conditioning that keeps both them comfortable and insulated bags at temperature in summer heat. A full-time driver burns 1-2 extra gallons of gas per day just idling, which at current prices eats $150-200 per month out of already thin earnings. This is not laziness — many restaurants have no indoor waiting area for drivers, and sitting in a 140-degree parked car in Phoenix or Houston is genuinely dangerous. The problem persists because delivery platforms externalize all vehicle costs to drivers and restaurants have no incentive to provide driver waiting areas since drivers are not their customers.

When identity theft victims place a fraud alert on their credit file, the alert merely instructs creditors to take "reasonable steps" to verify the applicant's identity before issuing credit. But there is no enforcement mechanism or penalty for creditors who skip this step, and many do because extra verification slows down their approval pipeline. A fraud alert does not block access to the credit report the way a freeze does, so a thief can still apply, get approved, and max out accounts in the victim's name while the alert is active. Victims who chose fraud alerts over freezes -- often because they were actively applying for credit themselves and could not freeze -- discover too late that the alert provided no actual protection. This persists because the FCRA defines fraud alerts as advisory rather than mandatory, and creditors face minimal liability for ignoring them since proving they failed to take "reasonable steps" requires expensive litigation.

Two companies, CDK Global and Reynolds and Reynolds, control the dealer management system market and have been sued for conspiring to overcharge dealers and block competitors from accessing dealer data. CDK alone paid $600 million to settle antitrust claims from 244 software vendors it allegedly locked out. This matters because every dealership's inventory, accounting, payroll, and customer data flows through these systems, so switching costs are enormous and dealers pass inflated software fees to consumers. The duopoly persists because OEM certification requirements effectively mandate one of these two systems, and both companies historically refused to share integration APIs with competing vendors.

American delivery platforms ask customers to set a tip amount before the driver even picks up the food, which means the tip is not a reward for good service but a bid for adequate service. Drivers see the total payout including tip before accepting, so low-tip or no-tip orders bounce between drivers for 30-60 minutes until someone desperate enough accepts, by which time the food is cold. Customers who tip well subsidize the system while no-tip customers get terrible service, creating a two-tier delivery experience that looks identical in the app's UI. Platforms designed pre-tipping because post-delivery tips were unreliable and drivers refused low-payout orders entirely, but the result is a system where the tip is functionally a bribe rather than a gratuity.

After a federal court vacated the CFPB rule banning medical debt from credit reports in July 2025, consumer protection fractured into a state-by-state patchwork. Fifteen states (including California, Colorado, and New York) passed their own bans, but the remaining 35 states still allow medical collections on credit reports. A consumer who visits an out-of-network ER while traveling can have a small balance sent to collections and reported to bureaus in their home state, even if the care was received in a state with protections. Many consumers wrongly assume all medical debt under $500 is now invisible to lenders, when the reality depends on which state's law applies, whether the debt was sold to a collector, and the timing of the collection report. This chaos persists because the federal court ruling created a regulatory vacuum, and credit bureaus default to reporting everything unless a specific state statute forces them not to.

When a dealership arranges financing, the lender provides a 'buy rate' but the dealer can silently add 2-3 percentage points as a 'dealer reserve' kickback, pocketing the spread. This costs American car buyers $25.8 billion in excess interest over the lives of their loans, and no law requires the dealer to disclose the markup or even tell you a lower rate was available. The system persists because lenders benefit from higher interest revenue they share with dealers, dealers are classified as arrangers rather than lenders so they dodge lending disclosure rules, and the consumer never sees the buy rate to know they are overpaying.

Restaurants pay 15-30% commission on every delivery order, so they raise their delivery menu prices to compensate — a $14 bowl in-store becomes $18 on DoorDash. Customers see the inflated price, add the delivery fee, service fee, and tip, and end up paying $28 for a $14 bowl, which makes them order less frequently and leave worse tips. Restaurants cannot show their real prices because the math does not work at 30% commission on thin margins, but customers blame the restaurant for being expensive rather than the platform for extracting the margin. This persists because restaurants are locked into a prisoner's dilemma: if they leave the platform, their competitors stay on it and capture all the delivery demand.

Specialty consumer reporting agencies that produce tenant screening reports scrape court records in bulk daily, but courts that seal or expunge eviction cases update their public records on a slower cadence -- sometimes weeks or months later. The result is that a tenant whose eviction was dismissed, sealed, or expunged still shows up as having an eviction record when a landlord runs a background check. Prospective tenants are denied housing based on stale, legally invalid data and often do not even know the screening report exists until after the rejection. Disputing these errors requires identifying which of dozens of obscure screening companies generated the report, then navigating their individual dispute processes. This persists because screening companies have no obligation to re-verify against court records after initial ingestion, and the FCRA's "maximum possible accuracy" standard is rarely enforced against them.

Finance and Insurance managers present extended warranties, GAP insurance, and paint protection as tiny monthly payment increases rather than total cost, making a $3,000 bundle sound like $12/month. This matters because 75% of buyers at some dealer groups end up with add-ons they never consciously agreed to, and those charges rolled into a 72-month loan at 9% APR balloon to over $4,500 by payoff. The practice persists because F&I is the single highest-margin department in a dealership, dealers face no legal obligation to present total add-on cost before signing, and the four-square worksheet deliberately obscures which line items changed.

Approximately 50 million Americans have no credit file at all, and another 62 million have "thin files" with too little history to generate a score. This disproportionately hits immigrants who had pristine credit histories in their home countries, young adults who have never borrowed, and low-income workers who pay rent and utilities on time but whose payments are invisible to the system. Without a score, these people cannot rent apartments (90% of landlords check credit), finance a car, or get a credit card, trapping them in a cycle where you need credit to get credit. The problem is structural: the credit system was designed around debt products, and the bureaus have no native mechanism to ingest positive payment data from rent, utilities, or subscriptions unless a third-party intermediary pays to report it.

There is no legal requirement for creditors to report to all three major credit bureaus, so many lenders, credit unions, and fintech companies report to only one or two. A consumer can have an 780 at Experian and a 710 at TransUnion simply because a well-managed credit card only appears on one report. Whether you get approved for a mortgage depends on which bureau the lender happens to pull, turning a supposedly objective financial profile into a lottery. Mortgage lenders pull all three and use the middle score, but auto lenders, landlords, and credit card issuers typically pull just one, and consumers have no way to know which one in advance. This persists because reporting is voluntary, bureaus compete for furnisher business rather than mandating comprehensive reporting, and no regulation requires furnishers to report symmetrically.

Between early and late 2025, Experian's rate of providing relief on consumer disputes plummeted from 20% to less than 1%, and TransUnion's declined by 50%, directly correlating with the Trump administration's dismantling of the CFPB. With the agency's enforcement staff cut from 248 to roughly 50 people and investigations frozen, the bureaus face no regulatory consequence for stonewalling disputes. Nearly 2.7 million complaints submitted to the CFPB went without any relief action. Consumers who previously had a credible escalation path now have no practical recourse short of hiring an FCRA attorney and suing. This structural collapse persists because credit bureaus are accountable to data furnishers who pay them, not to consumers whose data they sell, and the only counterweight -- federal enforcement -- has been deliberately neutralized.

A single commercial kitchen operates under multiple virtual brand names on delivery platforms — the same kitchen might appear as 'Mario's Italian,' 'Wing Boss,' and 'Burger District' — and customers have no way to know they are all the same facility cooking from the same line. When a customer gets food poisoning from 'Wing Boss' and avoids it, they unknowingly reorder from the same kitchen under 'Burger District' the next week. Health departments struggle to trace complaints because the business license is under a corporate entity name that matches none of the consumer-facing brands. Platforms allow this because more brand listings mean more search results and more commission revenue, and there is no regulatory requirement to disclose the physical kitchen behind a virtual brand.

Credit bureaus use probabilistic matching algorithms that consider a 7-out-of-9 digit SSN overlap a "complete match," causing two different people's credit files to merge into one. Fathers and sons with the same name, people with common Hispanic surnames, or anyone who shares a former address with a stranger can wake up to collections, late payments, and accounts they have never seen. Fixing a mixed file requires disputing each incorrect tradeline individually across all three bureaus, a process that can take months while the victim is denied mortgages or charged predatory interest rates. The problem persists because the bureaus prioritize ingesting billions of data points quickly over investing in deterministic identity resolution, and because FCRA lawsuits over mixed files are settled quietly without forcing systemic changes to the matching logic.

Customers who live in gated apartments or secure buildings type detailed access instructions into the delivery notes — gate code, building number, which elevator, which door — but drivers consistently ignore or miss them because the notes are buried below the fold in a tiny text field that requires scrolling past the map. The customer then gets a phone call from a frustrated driver sitting outside the gate, or worse, the driver marks the order as delivered and leaves it on the sidewalk outside. This is a UI problem that persists because delivery app product teams optimize the driver interface for map navigation and next-order speed, not for reading multi-line delivery instructions, and there is no structured field for access codes separate from general notes.

When consumers file credit report disputes with supporting documentation, the credit bureaus funnel everything through e-OSCAR, a system that compresses the dispute into a standardized 2-3 digit code from a dropdown menu. The furnisher (creditor or collector) receives only that code, not your letter, bank statements, or court documents. This means the "investigation" is just the furnisher re-checking its own records against a vague code and rubber-stamping the original data as verified. Consumers lose disputes not because they lack proof, but because their proof is architecturally excluded from the process. This persists because e-OSCAR is jointly owned by the three major bureaus, who have no incentive to build a more expensive system that would increase the rate of corrections against their paying data-furnisher clients.

When a restaurant notices a pricing error on their delivery platform menu — say a $12 entree accidentally listed at $2 — they cannot fix it instantly. Platforms like Grubhub and DoorDash require menu changes to go through a review queue that takes 24 to 72 hours, during which the restaurant must either fulfill orders at a massive loss or cancel them and tank their cancellation metrics. A single pricing typo during a busy weekend can cost a small restaurant hundreds of dollars in underpriced orders. This delay exists because platforms use centralized menu management teams to maintain formatting consistency and prevent fraud, but the system has no emergency override for obvious errors.

Delivery apps batch two or three orders from different restaurants into a single driver trip to reduce costs, but the second or third customer's food sits in the car for 20-40 extra minutes while earlier stops are completed. The second customer receives cold fries and congealed sauce, blames the restaurant, leaves a 1-star review, and the restaurant's rating drops for something entirely outside their control. Platforms keep stacking orders because paying one driver for two deliveries is dramatically cheaper than dispatching two drivers, and the cost of the occasional bad review is externalized onto the restaurant rather than absorbed by the platform.

Most delivery platforms hide the exact drop-off address from drivers until they accept the order, showing only an approximate distance or neighborhood. This means drivers unknowingly accept deliveries to apartment complexes with no gate codes, high-rise buildings with broken elevators, or military bases with 15-minute security checks — turning a seemingly quick delivery into a 30-minute ordeal that pays the same $3.50. Drivers who cancel after seeing the real address get hit with completion rate penalties that can lead to deactivation, so they are trapped. The platforms hide addresses intentionally because if drivers could cherry-pick, hard-to-reach customers would never get deliveries, so the information asymmetry is a deliberate design choice to distribute undesirable orders.

Delivery platforms like DoorDash and Uber Eats ship dedicated tablets to restaurants, but these Android tablets auto-install OS and app updates during peak hours, causing the device to freeze or reboot while orders are streaming in. When the tablet goes down for even 3 minutes during a Friday dinner rush, the restaurant misses incoming orders, those orders get reassigned or canceled, and the restaurant's acceptance rate drops — which the algorithm then punishes by deprioritizing them in search results for days afterward. This persists because delivery platforms control the tablet firmware and push updates on their own schedule, and restaurants have zero ability to defer or schedule updates since the devices are locked down.