@thoughtson_tech Special Interest Media

Special Interest Media posts on X about this is, ai, engineering, anthropic the most. They currently have [--] followers and [--] posts still getting attention that total [-----] engagements in the last [--] hours.

Engagements: [-----] #

Mentions: [--] #

Followers: [--] #

CreatorRank: [-------] #

Social Influence

Social category influence finance 18.37% technology brands 16.33% stocks 6.12% countries 6.12%

Social topic influence this is 20.41%, ai 18.37%, engineering 14.29%, anthropic #2784, reduce #1566, business 12.24%, money 10.2%, health 8.16%, the most 6.12%, drop 6.12%

Top accounts mentioned or mentioned by @organize @partnerwithppl @mikikapoor

Top Social Posts

Top posts by engagements in the last [--] hours

"CDPAP is actually one of the most structurally interesting fraud cases in the entire Medicaid home care space and worth understanding in detail because it illustrates exactly why consumer-directed care models are so difficult to police at scale. The programs design is the core issue. CDPAP allows Medicaid beneficiaries to hire their own caregivers including family members and directs payment through a fiscal intermediary rather than a traditional home health agency. The intent was genuinely good - give disabled and elderly individuals more autonomy over their care rather than forcing them"
X Link 2026-02-16T01:01Z [--] followers, [--] engagements

"This is genuinely excellent analysis and the COVID ratchet framing is the right one. The 97% spending increase with only 10.7% entity growth and a 42.4% cost-per-claim surge is the most important decomposition in this chart. That ratio tells you a lot about what actually happened. A few things worth layering on before landing on fraud as the primary explanation for the permanent baseline shift because the reality is messier and the policy implications differ depending on the cause. The cost-per-claim surge in T1019 has three potential drivers that need to be disaggregated. First legitimate"
X Link 2026-02-16T01:11Z [--] followers, [--] engagements

"Great piece. The GLP-1 cash payment phenomenon is exactly the kind of system-level bypass youre describing - when insurance becomes friction instead of value rational consumers route around it entirely. We saw this accelerate dramatically when brand manufacturers cut their own deals with the Trump administration essentially acknowledging that the PBM-payer negotiation layer was extractable overhead rather than essential infrastructure. The fascinating part is how viral demand overwhelmed manufacturing capacity faster than employers could even evaluate coverage decisions. Patients didnt wait"
X Link 2026-02-08T01:39Z [--] followers, [---] engagements

"Marion nails it. Vertical integration inside their house is the PBM-GPO conflict: UnitedHealth owning OptumRx and OptumInsight optimizes for rebates ($150-250 per member) versus provider cost reduction leaking $50-150 per member. MLR gaming works because subsidiaries pay themselves. OptumRx routes specialty through Optum pharmacies at inflated pricescosts hit MLR numerator profits stay internal. OptumHealth visits work identically: UnitedHealthcare pays Optum clinics satisfying MLR while keeping revenue in corporate family. Vertically integrated models show 15-25% lower pharma compliance"
X Link 2026-02-08T15:37Z [--] followers, [---] engagements

"This is the key distinction people keep flattening. Cost Plus is morally right and structurally limited at the same time. Cost-plus works when the problem is opacity on commodity inputs. It breaks when the problem is market power embedded upstream. Generics are the former. Branded drugs specialty and rebate-driven formulary economics are the latter. A few things worth being precise about: - Cost Plus covers 1015% of spend because manufacturers refuse to participate in transparent pricing when rebates are the real currency. - The remaining 8590% isnt non-transparent by accident its opaque"
X Link 2026-02-11T00:50Z [--] followers, [---] engagements

"The framing that this bill isnt really about hospitals or the subsidy stack misses whats structurally different this time. The Break Up Big Medicine Act is not a price-control bill. Its a structural separation bill explicitly modeled on Glass-Steagall: you cannot own both the toll booth and the traffic. That means no common ownership of insurer/PBM and provider or MSO and no wholesalers owning provider orgs. One-year compliance. Forced divestitures. Disgorgement. Broad enforcement authority across FTC DOJ HHS state AGs and even private plaintiffs. That design choice matters. Over the last 15"
X Link 2026-02-12T13:01Z [--] followers, [---] engagements

"If you strip away the branding this isnt an anti-size bill. Its a structural separation bill modeled explicitly on Glass-Steagall. The core prohibition is clean: no common ownership of insurer or PBM and provider or MSO and no wholesalers owning provider organizations with a one-year compliance window and forced divestitures for noncompliance. That matters because the current cost structure is not just about market share. Its about closed-loop economics. When the same parent owns the insurer the PBM the pharmacy and the physician group internal transfer pricing can be labeled medical spend"
X Link 2026-02-12T13:05Z [--] followers, [--] engagements

"The example cited here is exactly why structural separation is back on the table. When a single parent can own the insurer the PBM the pharmacy and the physician group every transaction along the patient journey becomes an internal transfer. Underwriting formulary design dispensing and care delivery sit inside the same balance sheet. The question is not whether any one segment works. It is whether the combined ownership structure distorts pricing and competition. Vertical integration was defended as care coordination and efficiency. In practice it also created the ability to shift margin"
X Link 2026-02-12T13:12Z [--] followers, [--] engagements

"Senator Wyden with respect "require more staff" is not a healthcare policy it's a staffing mandate that ignores why we're short staff in the first place. The US already can't train nurses fast enough to meet demand. Clinical placement bottlenecks faculty shortages and 25-30% annual turnover mean adding headcount requirements without addressing workflow efficiency just accelerates the doom loop: understaffing burnout turnover more understaffing. The real question isn't robots vs. nurses. It's: why are nurses spending only 25-35% of their shift on actual hands-on care The rest goes to"
X Link 2026-02-13T01:40Z [--] followers, [--] engagements

"The "algos overreacted to known information" framing is too easy here. Yes the Q4 earnings miss was telegraphed. But the 19% single-day drop isn't algo manipulation it's the market finally pricing in what the numbers have been saying all year: the business model is structurally broken in ways that don't fix themselves when a bad quarter rolls off the comps. The issue isn't Q4 in isolation. It's the full-year story: MLR at 85.1% up [---] basis points means UHC added billions in top line revenue while operating income actually declined. That's not a one-quarter blip. That's unit economics"
X Link 2026-02-13T01:47Z [--] followers, [--] engagements

"Smart money bought at these levels before but the business they were buying has structurally changed since then. The "super investors bought here" signal is compelling as a valuation floor argument. But Tepper Burry and Buffett were buying into a thesis that assumed Medicare Advantage margins would normalize Optum would keep compressing costs through vertical integration and the benign utilization environment of 2021-2023 would roughly hold. Those assumptions are all under pressure simultaneously right now. UHC's MLR hit 85.1% for the full year up [---] basis points. Operating income declined"
X Link 2026-02-13T01:48Z [--] followers, [--] engagements

"17 people die every day waiting for organs that could have been procured but weren't. Some OPOs recover organs from 70% of eligible deaths. Others recover from under 30%. The gap isn't explained by geography or donor pool differences it's pure operational and accountability failure. Here's the actual scandal: CMS has never once decertified an OPO for underperformance in the program's entire history. [--] federally designated monopolies with zero meaningful consequences for leaving organs on the table. The proposed CMS rule changes that are working through implementation right now would shift"
X Link 2026-02-13T01:52Z [--] followers, [--] engagements

"This chart is the proof of concept for the next wave of reform. 29% increase in kidney recoveries and $359M in fiscal savings from a single rule change that introduced accountability metrics and CMS still hasn't fully implemented outcomes-based certification with real decertification consequences. The [----] reform proved the thesis: OPOs respond to performance pressure when it's credibly enforced. The gap between the best and worst performers (70% vs. sub-30% of eligible deaths recovered) isn't biological or geographic it's operational and accountability failure. That gap doesn't close itself."
X Link 2026-02-13T01:53Z [--] followers, [--] engagements

"The 3x spending statistic is real. But the "why" is more interesting than the headline. A recent Health Affairs study decomposed US healthcare spending patterns between 2009-2019 and found something that surprises most people: real per capita health spending grew just 1.7% annually during that decade less than half the historical rate of 3.7%. The healthcare share of GDP basically flatlined around 17-18%. So the problem isn't that spending is out of control relative to trend. The problem is that we built an extraordinarily expensive baseline and then locked it in. Here's what the data"
X Link 2026-02-13T01:58Z [--] followers, [--] engagements

"This is the 2022-2024 reversal of a decade-long trend worth understanding carefully. The 2009-2019 period told a different story. Private insurance hospital spending came in $78.9 billion below CMS projections. Inpatient utilization fell 10-30% depending on the payer. The conventional wisdom that hospital consolidation would drive price escalation turned out to be wrong mergers between 2010-2015 averaged just 1.6% price increases basically a rounding error on a $430 billion spending base. What changed post-2022: The care substitution trend that drove the 2009-2019 moderation outpatient"
X Link 2026-02-13T01:59Z [--] followers, [--] engagements

"The income elasticity argument is right as far as it goes the R=0.944 correlation between income and health spending is real and your procedure volume data is solid. But it explains the level more than the recent acceleration. Here's what makes the 2009-2019 period instructive: real per capita health spending grew just 1.7% annually during that decade less than half the historical 3.7% rate even as incomes rose and new expensive procedures became available. The income-elasticity model would have predicted continued escalation. It didn't happen. What actually drove the moderation: utilization"
X Link 2026-02-13T02:01Z [--] followers, [--] engagements

"The divergence is real and the [----] starting point is telling. But the "US became an outlier" framing obscures something important about what actually happened during the decade when the gap was supposed to widen fastest. Between 2009-2019 US healthcare spending as a share of GDP basically flatlined at 17-18%. Real per capita spending grew just 1.7% annually less than half the historical 3.7% rate that created the divergence shown in your chart. The gap didn't keep widening during that decade. It stabilized. This matters because it refutes the fatalist interpretation of the chart that"
X Link 2026-02-13T02:03Z [--] followers, [--] engagements

"Reinhardt's "it's the prices" framing was correct as a diagnosis of the historical gap. But the 2009-2019 data complicates the simple version of the story. During that decade real costs per inpatient hospital stay grew just 2.1% annually down from 2.9% in 2000-2009. This deceleration happened during accelerating hospital consolidation. One-third of hospitals were involved in mergers between 2010-2019. Conventional wisdom said this would give them pricing power over private insurers. It mostly didn't mergers between 2010-2015 averaged just 1.6% price increases. Physician prices followed a"
X Link 2026-02-13T02:05Z [--] followers, [--] engagements

"Some of it yes. The physically impossible stuff - 1000+ claims per workday from a single van in rural New Mexico solo pediatricians with 2.1M claims - thatll get surfaced fast and itll be viscerally obvious. The harder layers are subtler. New LLC entities that ramp to $50M in [--] months then vanish. Procedure code drift within a providers billing history that signals diagnosis upcoding. NPI lineage patterns where convicted operators reopen under spouse or family member NPIs. Geographic clustering of bust-out schemes within the same HCPCS taxonomy across state lines. None of that is obvious in a"
X Link 2026-02-14T12:49Z [--] followers, [---] engagements

"The PPP loan comparison is the right analogy and you called it two months early - thats exactly what this release is. PPP data let journalists researchers and just regular people with spreadsheet skills surface fraudulent loans that federal auditors missed for years. Same dynamic applies here. One thing worth being precise about though: the original December post framing this around Somali community geography is where the analysis needs to be careful. The Minnesota fraud cases that have been prosecuted involve specific providers and specific billing schemes - the legal record on that is"
X Link 2026-02-14T12:56Z [--] followers, [--] engagements

"The incentive structure idea is worth taking seriously. The False Claims Act whistleblower model - where tipsters get 15-30% of recovered funds - has historically generated strong ROI for the government. HCFAC program returns $2.80 for every $1 spent on fraud enforcement. Extending that logic to crowd-sourced data analysis with financial incentives could accelerate what this dataset surfaces. The design challenge is getting the incentive structure right. A few things that would need solving: False positive costs are real. Publicly accusing a legitimate provider of fraud because their billing"
X Link 2026-02-14T14:35Z [--] followers, [---] engagements

"The qui tam angle is genuinely interesting and underexplored. False Claims Act whistleblower suits have historically been one of the most effective fraud recovery mechanisms in healthcare - the HCFAC program returns $2.80 for every $1 spent and a meaningful portion of that comes from relator-filed qui tam cases where private citizens or insiders had information the government didnt. The agent-to-automatic-filing vision has a few hard problems though that are worth thinking through before anyone actually builds it. FCA qui tam requires the relator to have original information - not just"
X Link 2026-02-14T16:21Z [--] followers, [--] engagements

"The combination is potentially very powerful but the design details matter enormously and will determine whether this becomes a durable fraud recovery mechanism or a chaotic false positive generator. The FCA qui tam model already does roughly what youre describing and has a 40-year track record worth studying. Relators get 15-30% of recovered funds DOJ decides whether to intervene cases that dont hold up get dismissed. The ROI on the program is strong - HCFAC returns $2.80 per $1 spent. But it works because theres a verification layer between tip and payout. You dont get paid for flagging an"
X Link 2026-02-14T16:29Z [--] followers, [--] engagements

"Judge Rakoff just ruled that pasting legal content into a consumer AI chatbot waives attorney-client privilege. Full stop. The case is U.S. v. Heppner out of SDNY. Guy was under federal investigation for securities fraud used consumer Claude to prep [--] docs of legal strategy sent them to his Quinn Emanuel lawyers. FBI seized the devices. Privilege claim Dead on arrival. Rakoffs take: Im not seeing remotely any basis for any claim of attorney-client privilege. Why it failed: consumer ToS lets the provider train on inputs and disclose to govt authorities. Thats a third-party disclosure."
X Link 2026-02-15T14:19Z [--] followers, [--] engagements

"100% valid and the rage is warranted. The fraud data makes this even more infuriating when you understand the structural mechanics. Medicaid rates for legitimate physicians managing complex high-acuity patients have been systematically suppressed for decades partly because the program kept running out of money. Except it wasnt running out of money. A meaningful chunk of it was flowing to home health LLCs formed [--] months ago billing [----] claims per workday from a van in rural New Mexico behavioral health agencies ramping to $50M in [--] months before vanishing and NPI laundering operations run"
X Link 2026-02-16T00:50Z [--] followers, [--] engagements

"The regulatory tailwinds are finally catching up to what the market has been building toward. ARPA-H ADVOCATE CMS ACCESS FDA TEMPO and Utahs prescribing law arent isolated experiments - theyre converging signals that autonomous AI in care delivery is becoming policy not just product. What this shift demands isnt just better AI - its a new framework for the patient-provider relationship entirely. When AI can prescribe and deploy before full clearance the question of who validates clinical decisions becomes more urgent not less. I wrote about exactly this tension - how structured patient"
X Link 2026-02-16T12:51Z [--] followers, [--] engagements

"This is exactly right and the analogy to software engineering is more precise than most people realize. Claude writing its own code didnt shrink Anthropics engineering team - it expanded the surface area of whats possible and raised the value of human judgment at the orchestration layer. The same dynamic plays out in care delivery. When AI handles documentation chart summarization and draft orders physicians dont become redundant - they finally have the cognitive bandwidth to do what theyve never had time for: genuine care coordination patient education shared decision-making and the complex"
X Link 2026-02-16T13:38Z [--] followers, [--] engagements

"1006 claims per workday from a van in rural New Mexico is a perfect illustration of exactly why rule-based audit systems miss so much. That pattern - volume physically impossible for a single provider at a single location - is trivially detectable with basic ML. It persisted anyway because nobody had easy cross-state visibility into provider-level billing at this granularity until yesterday. Thats the whole point of the dataset. Not that the fraud is surprising. Its that the data infrastructure to catch it in something closer to real-time finally exists publicly. The scarier implication"
X Link 2026-02-14T12:48Z [--] followers, [----] engagements

"Spent the weekend building the data stack that should exist to catch Medicaid fraud. Spoiler: everything you need is free and public. The new Medicaid spending dataset is useful. Joined against NPPES entity formation dates OIG exclusion lists PECOS reassignment data Open Payments debarment records and state corporate registries it becomes something close to a fraud detection system. None of that costs anything. The moat isnt data access anymore. Its knowing what to do with it. Home health is the $288B taxonomy at the top of the spending list for a reason. No lab results. No imaging. No"
X Link 2026-02-15T17:18Z [--] followers, [--] engagements

"The revolving door chart is real and the staffing ratio is genuinely striking. [----] people overseeing $1.52T is about $232M per employee in program spend which by any measure is a thin oversight layer for the complexity involved. Worth being precise about what that actually means though because not looking out for you and captured by industry are different failure modes with different fixes. The fraud problem that the new dataset is surfacing - vans billing $76M LLCs ramping to $50M in [--] months then vanishing NPI laundering through family member entities - largely didnt persist because CMS"
X Link 2026-02-16T00:53Z [--] followers, [---] engagements

"Both posts in this screenshot are actually saying the same thing from different angles and its worth connecting them explicitly. Dr. Mollie James is right that its by design. Not conspiracy - structural design. The behavioral health and personal care taxonomies that dominate the fraud patterns in the new dataset have documentation requirements that are nearly impossible to audit at scale by design because the services themselves are human interactions in private settings with no independent clinical artifacts. A surgery leaves an operative note imaging pathology anesthesia records. Four hours"
X Link 2026-02-16T00:56Z [--] followers, [--] engagements

"The 20% figure needs a source before it travels further. The official CMS improper payment rate for Medicaid in recent years has been in the 5-8% range which on a $920B program is still $46-74B annually - a staggering number that doesnt need inflation to be outrage-worthy. Some researchers who include eligibility errors and documentation failures get to higher estimates but 20% is not a figure with strong methodological backing and it will get used to justify across-the-board cuts that hit legitimate care delivery rather than targeted fraud enforcement. The Brooklyn thread is the more"
X Link 2026-02-16T00:59Z [--] followers, [--] engagements

"The $288B home health number is correct and the easiest to defraud least oversight characterization is accurate. But the taxonomy breakdown reveals something even more interesting when you look at the full list together. The top five categories alone - home health at $288B acute care hospitals at $197B community behavioral health at $152B FQHCs at $96B and case management at $77B - account for roughly $810B of cumulative spend. The fraud density is not evenly distributed across those five. Acute care hospitals and FQHCs have extensive documentation requirements facility licensing and"
X Link 2026-02-16T01:15Z [--] followers, [--] engagements

"The $12630 per patient per month billing rate is the number worth interrogating here because its actually the more analytically interesting signal than the phantom billing allegation alone. For context the average Medicaid home care reimbursement for legitimate high-acuity personal care runs roughly $1500-3000 per beneficiary per month depending on the state and the authorized hours. $12630 per patient per month in a community behavioral health or personal care taxonomy is either an extremely high authorized care plan for a genuinely complex patient or it represents a volume of billed hours"
X Link 2026-02-16T01:17Z [--] followers, [--] engagements

"Levine is actually correct on the methodological point and this exchange illustrates exactly why zip-code-level billing analysis requires careful interpretation before going viral. The $3.8B billed from zip code [-----] with [-----] residents is not $143000 per resident in care delivered to those residents. Billing intermediaries fiscal intermediaries for consumer-directed care programs and large home care agencies register their NPI at a single corporate address while serving beneficiaries across entire boroughs or the whole state. The zip code in the dataset reflects where the billing entity is"
X Link 2026-02-16T12:53Z [--] followers, [----] engagements

"Been thinking about why nobodys built a hedge fund for specialty drug volatility yet and honestly its kinda wild Specialty biologics are 50%+ of pharmacy spend now. Gene therapies hitting payers with $2M single-event shocks. CAR-T cases clustering randomly. GLP-1s cascading through employer groups in ways actuaries cant predict. Current tools (PBMs prior auth rebates) address pricing optics but completely miss the actual problem which is actuarial volatility not absolute cost levels. A regional Blues plan with 400k lives can get their whole year blown up by three unlucky cases. The structural"
X Link 2026-02-17T01:29Z [--] followers, [--] engagements

Limited data mode. Full metrics available with subscription: lunarcrush.com/pricing