Free Daily Podcast Summary
The Innovation Attorney Podcast: Daily Summaries Delivered
by The Innovation Attorney
The Innovation Attorney is a publication about the intersection of venture capital, technology, public policy and law. theinnovationattorney.substack.com
Latest Episodes
The most recent episodes — sign up to get AI-powered summaries of each one.
- Today36 min
Usage Based Pricing for AI Products
Salesforce, Zendesk, and HubSpot shifted AI pricing toward usage based and outcome based models during 2025 and 2026 because seat based subscriptions could not track token costs or delivered value. A venture funded AI product should move to usage based or hybrid pricing when its inference cost per customer varies by more than roughly three to one across the user base, the threshold several pricing operators use to flag margin risk under flat seat pricing. Salesforce priced Agentforce at $2.00 per conversation at its October 2024 launch, then layered in $0.10 per action Flex Credits in May 2025, and still sells seats starting at $125 per user per month under its Agentic Enterprise License Agreement, because no single model fits every buyer. The right answer depends on gross margin at the ninetieth percentile user, deal size, and how predictably the company can forecast consumption, not on which model draws the most attention.The underlying driver is a token cost collapse that moved faster than most pricing committees tracked it. Input token prices for frontier grade language models fell from $30 per million tokens for GPT-4 in March 2023 to roughly $0.10 per million tokens for comparable capability by early 2026, a decline close to three hundred times in under three years. Anthropic cut its own 2025 gross margin forecast to about 40 percent as inference costs outran its revenue growth, even as 2025 revenue reached approximately $10 billion. A company billing a flat $29 monthly seat against a ninetieth percentile user consuming $21.00 in tokens has already lost money on that account before any other cost is counted.The comparative evidence favors a hybrid structure over either pure model standing alone. Vertice’s second quarter 2026 contract data placed usage based pricing as the single most common primary model at 36.5 percent of software contracts, ahead of per user pricing at 32.4 percent, while L.E.K. Consulting’s April 2025 valuation analysis found hybrid pricing companies carrying a 6.9x mean revenue multiple against 6.5x for pure subscription companies. For an investor, the practical implication is that the pricing model a portfolio company selects in 2026 shows up directly in gross margin, net revenue retention, and the multiple the company commands at its next financing, before any product feature does.Definition, Scope, and Technical ContextUsage based pricing charges a customer for the metered quantity of a resource consumed, most commonly the input and output tokens processed by a large language model, while seat based pricing charges a fixed fee per named user regardless of how much of the product that user consumes. This analysis treats outcome based pricing as a third category: a charge triggered by a completed result, such as HubSpot’s $0.50 fee per resolved Breeze Customer Agent conversation after April 14, 2026, or Salesforce’s $2.00 per Agentforce conversation at its October 2024 launch. The analysis covers venture funded, cloud delivered AI products sold to businesses, meaning software companies that purchase frontier model access through an application programming interface rather than companies that train and sell their own foundation models.The analysis excludes consumer facing subscription products such as a standalone chatbot application, because consumer willingness to pay follows different psychology than enterprise procurement, and it excludes open source, self hosted deployments, where the customer bears the inference cost directly and the vendor pricing question does not arise in the same form.The relevant engineering classification is metered cloud services pricing applied to a workload whose unit cost is itself falling rapidly and whose consumption per user varies far more than consumption of a conventional software seat. A seat on a conventional SaaS product, a project management tool or a word processor, consumes a roughly fixed quantity of server compute regardless of how actively the named user works, because the software is deterministic and the marginal cost of an additional click is near zero. An AI agent seat consumes a variable quantity of token generating compute tied directly to how many queries the user submits and how complex each query is, so the cost basis underneath the seat is no longer flat. That difference, not any feature of the product itself, is what forces the pricing question addressed throughout this report.Underlying Scientific and Engineering PrinciplesA large language model generates output one token at a time through autoregressive decoding, and the compute cost of that process is driven by the number of parameters activated per token, the length of the input context the model must attend to, and the hardware generation running the calculation, which is why token price is the correct unit of cost measurement rather than a flat per query or per seat charge. Every additional output token requi
- Yesterday14 min
Satellite Close Approach Defense
A radar imaging satellite needs 5 to 10 meters per second of reserve, onboard precise navigation, and relative sensing to answer a close approach like COSMOS 2612.¹ That figure comes from first order orbital mechanics applied to the October 2026 encounter, in which the Russian satellite COSMOS 2612 closed from about 400 kilometers to roughly 500 meters of ICEYE X36, a 90 kilogram synthetic aperture radar satellite registered in the United States.¹²The encounter ran from October 15 to October 19, 2026. Public orbital elements show COSMOS 2612 spent about 67 meters per second of change in speed after September 30 to arrive, and about 105 to 117 meters per second in May to match the orbital plane of its target.² A radar satellite operator cannot match that budget. A reserve of 5 to 10 meters per second buys a separation of 85 to 170 kilometers per orbit after a coplanar lowering burn, which exceeds the roughly 10 kilometer standoff that states have generally kept.⁴Propellant is half the answer. Public element sets for small objects in this orbit carry in track uncertainty near 1.9 kilometers, which is nearly four times the distance LeoLabs reported at closest approach.² A satellite that depends on public tracking data cannot see an event of this size. It needs its own satellite navigation receiver, a way to sense a neighbor directly, and a software path that turns a detection into a burn within one orbit.The legal exposure follows the engineering. Operators licensed under 47 CFR 25.114(d)(14) already disclose collision avoidance capability, and the Liability Convention of 1972 assigns fault based liability for damage in space.⁷¹⁰ Neither text addresses a deliberate approach that causes no collision.Detailed FindingsWhat Happened Between COSMOS 2612 and ICEYE X36?COSMOS 2612 made a controlled, multi week approach to ICEYE X36 between September 30 and October 19, 2026, and held position within about 500 meters of the target on October 17 and 18, according to the commercial tracking firm LeoLabs.¹ The Russian satellite was one of six, COSMOS 2609 through 2614, launched on a single Soyuz 2.1b from Plesetsk in April 2026.³Public element sets show the approach began with phasing, not speed. COSMOS 2612 raised its orbit about 47 kilometers on September 30, descended 47.8 kilometers on October 11 and 12 to match the target altitude, and closed from about 480 kilometers to about 3 kilometers on October 15 and 16 with burns of under 2 meters per second each.² It then held station behind the target on a loop about 1.6 kilometers tall, and on October 18 raised its orbit 2.3 kilometers to leave. By October 19 it was 292 kilometers behind and drifting away at about 350 kilometers per day.²The closest distance carries uncertainty. LeoLabs reported about 500 meters with high confidence. An independent reading of the public catalog supports a range of roughly 150 to 650 meters in the early hours of October 18, and a 25 meter figure that circulated cannot be derived from public data at all.²³ ICEYE reported moving its own satellite and told the press its constellation architecture is designed for continuity.¹² The Pentagon declined to comment.⁴How Much Propellant Does a Radar Satellite Need to Respond to a Close Approach?A radar imaging satellite needs a dedicated avoidance reserve of 5 to 10 meters per second, sized for a coplanar phasing burn instead of a plane change. These figures are the author’s first order calculation, using a circular orbit near 500 kilometers, an orbital speed of about 7.6 kilometers per second, and a semimajor axis of about 6,871 kilometers. The altitude is an assumption drawn from published descriptions of the ICEYE constellation.⁶The calculation reproduces the Russian numbers, which is the reason to trust it. A two burn altitude change of Δa costs a total change in speed near v times Δa divided by twice the semimajor axis. For the 47.8 kilometer descent that gives about 26 meters per second, and public estimates put the real burn near 27 to 28.² A plane change of 0.88 degrees costs about v times the angle in radians, or about 117 meters per second, which sits at the top of the 105 to 117 meter per second range estimated for the May burns.²Applied to the defender, the same equations give the exchange rate that matters. ICEYE lowered its satellite about 0.5 kilometers on June 25 and 26, 2026, a maneuver under 1 meter per second, and the calculation puts that burn near 0.3 meters per second.² A 5 meter per second reserve raises the altitude change to about 9 kilometers. An altitude offset produces an along track drift of roughly three times pi times the offset per orbit, so a 9 kilometer offset opens about 85 kilometers per orbit, and a 10 meter per second reserve opens about 170 kilometers.Plane changes lie outside any defender’s budget. Moving the orbital plane by even 0.1 degree costs about 13 meters per se
- 2 days ago40 min
Rare Earth Metals
Antimony refining startups selling to defense buyers need ASTM B237 Grade A metal at 99.80 percent antimony or better, with lot level assays and mine origin records for every ingot.American Tungsten and Antimony reported on October 13, 2026 that one ingot from its Del Sol refinery in Amargosa Valley, Nevada assayed 99.8 percent antimony, while four earlier ingots assayed 99.5 to 99.6 percent. One passing sample clears the Grade A floor and shows a process change worked. It shows no repeatability, and the Defense Logistics Agency and the Army buy lots, records and continuity of supply. Financeable offtake terms include a stated minimum quantity, a price floor, an agreed assay protocol with a named umpire laboratory, and acceptance periods short enough to protect working capital.Three domestic projects frame the supplier side: a Defense Logistics Agency contract worth up to $245 million for ingots from United States Antimony Corporation, a $2.9 billion Export Import Bank loan to Perpetua Resources for its Idaho mine, and Nova Minerals’ Alaska plant, where all key equipment arrived on October 6, 2026.The Grade A threshold, the federal awards and the July 20, 2026 tracing order for defense contractors are established. Repeatability at Del Sol, completion of its parent’s acquisition of the site, and the price the government pays when spot prices fall are not.Detailed FindingsWhat Did the Del Sol Assay Actually Show?The Del Sol result is one 99.8 percent antimony assay on one ingot, produced after the company changed its electrowinning conditions, and it sits exactly at the ASTM B237 Grade A minimum of 99.80 percent. In its October 13, 2026 announcement the company said NSL Analytical Services of Cleveland, Ohio tested drill shavings from five ingots made from ore from the White Spar Antimony Mine in Yavapai County, Arizona. Sample 294 of the S260930 series returned 99.8 percent. Samples 290 through 293 returned 99.5 to 99.6 percent, and only the fifth ingot came from the optimized electrowinning conditions.The laboratory measured impurities by inductively coupled plasma mass spectrometry, sulfur by pyrohydrolysis followed by ion chromatography, and mercury by cold vapor atomic absorption. It calculated antimony by difference. A by difference result equals 100 percent minus every impurity measured, so an element nobody measured counts as antimony. A buyer’s specification therefore has to name the element list as well as the antimony floor. Reporting to one decimal place also leaves no margin: a true value of 99.76 percent rounds to 99.8, so the rounding rule belongs in the contract.Coverage of the announcement states that the 99.8 percent sample met the Grade A chemical limits for lead, arsenic, copper, nickel and iron. The coverage reviewed published no impurity values, no recovery rates, no volumes, no funding and no offtake commitment, and it records that the company’s acquisition of the entities holding Del Sol and White Spar has not completed. The company’s first ingot announcement came on September 22, 2026 and said the ingot was made without new processing infrastructure or new permitting for the current campaign. The company now says the result lets it approach the Defense Logistics Agency, Project Vault and defense primes. It also says it expects consistent output above the commercial grade, which it has not demonstrated.On October 5, 2026 a company executive joined a delegation at White House and Office of the United States Trade Representative roundtables, where the contacts included the Trade Representative, a National Security Council official and the director of the Office of Strategic Capital. The discussion centered on funding and offtake with the federal government and defense primes. No agreement was reported.What Grades and Impurity Limits Do Antimony Buyers Use?Antimony ingot trades against three reference points: a 99.65 percent commercial grade, a 99.85 percent premium grade, and ASTM B237 Grade A at 99.80 percent minimum. The standard covers refined antimony as ingot, pig or cake in Grades A and B, sets limits for arsenic, sulfur, lead and other impurities, and allows any smelting or refining route, from ore or recycled feed, that meets the chemistry. Reports of the standard give Grade B a 99.50 percent minimum.The 99.65 percent commercial grade, sold as Regulus II or Grade II, carries maximums of 0.15 percent arsenic, 0.005 percent selenium and 0.06 percent sulfur in one published trading specification. A supplier data sheet adds 0.20 percent lead, 0.002 percent copper and 0.02 percent iron for the same antimony floor. A 99.85 percent ingot assessed by a price reporting agency carries 0.05 percent maximum arsenic, 0.15 percent maximum lead and 100 parts per million maximum bismuth, and the supplier data sheet lists 0.085 percent lead, 0.001 percent copper and 0.04 percent sulfur for its 99.85 percent pro
- 3 days ago33 min
Redefining the Product Manager Role
An AI coding agent such as Claude Code now lets a single product manager draft a specification, synthesize customer interviews, build a working prototype, and in a growing number of companies open a pull request that ships to production, work that as recently as 2024 moved through an engineer by default. Reporting published June 27, 2026 described Anthropic’s own growth team being told to hire more product managers rather than fewer once its engineering organization began shipping at roughly three times its prior headcount, and the same reporting put the traditional one to eight product manager to engineer ratio at an effective one to twenty inside agent heavy teams. The shift is uneven. A 2024 Pragmatic Institute survey found that sixty five percent of product professionals had already integrated AI into their work, concentrated in routine and descriptive tasks, while a 2026 developer survey found that forty six percent of engineers still distrust the output AI tools produce, up from thirty one percent the year before.The central finding of this report is that the handoff point between product management and engineering has moved, not disappeared. Product managers at Amazon, Google, Meta, LinkedIn, and a small cluster of independent software companies are now shipping prototypes, internal tools, and in some documented cases production fixes directly, while engineers retain sign off authority on anything that touches authentication, data integrity, or system architecture. A March 18, 2026 content analysis of 638 published Lenny’s Newsletter and podcast items, built with a Claude Opus 4.6 multi agent research team, found that artificial intelligence became the dominant lens for discussing product management starting in the second quarter of 2024, rising from under four percent of coverage in late 2022 to sixty seven percent in the first quarter of 2026.For investors and operators allocating capital to product organizations, the practical implication is that headcount models built on a fixed product manager to engineer ratio are already out of date at companies that have adopted agentic coding tools, and the associate product manager track, the traditional entry point for junior hires, is contracting faster than any other segment of the profession.Definition, Scope, and Technical ContextAn AI agent, for purposes of this report, is a large language model deployed with tool access: the ability to read and write files, execute code, run tests, query databases, browse the web, and complete multi step tasks with limited human prompting between steps. This definition separates an agent from a conversational chatbot. Tal Raviv, writing for Lenny’s Newsletter on April 29, 2025, drew this same line, distinguishing agents by their capacity to act proactively, draw on company context and live data, take real actions, and check their own output. Claude Code, Anthropic’s command line coding agent, and comparable tools such as Cursor and OpenAI’s Codex, are the dominant instances of this category inside product organizations as of October 2026.The scope of this report is the set of tasks that moved, or are moving, from a product manager’s engineering handoff to an agent: drafting specifications, synthesizing user research, building interactive prototypes, writing code for internal tools and small production fixes, and the review structure that now governs that code before it reaches customers. The report excludes the law and regulation that may someday govern agent authored software, consistent with the scope the user set for this engagement, and it excludes agent use inside pure engineering teams that does not touch the product management function.The system under analysis sits inside the broader category of agentic software engineering, a 2025 and 2026 development that layers planning, tool use, and self checking on top of a foundation model’s text generation capability. The product management variant of this system differs from its engineering counterpart principally in who holds the keyboard: the operator is frequently a person without a formal computer science background, supervising an agent that performs the implementation work a software engineer once performed by hand.Underlying Scientific and Engineering PrinciplesAn agentic coding tool operates as a loop: the model receives an instruction, forms a plan, calls a tool such as a shell command or a file edit, observes the result, and repeats until it judges the task complete or asks for human input. Paweł Huryn, writing in The Product Compass on May 31, 2026, described building three public repositories this way, including a Visual Studio Code extension called PM Brain supported by more than eight hundred automated tests, without personally reading most of the underlying code. The mechanism that makes this possible is a combination of a large context window, the ability to execute and observe code and run it
- 4 days ago1h 13m
Soviet Space Reactor Design, Recovery Cost, and Orbital Debris
The BES 5 Buk reactor, a 100 kilowatt thermal Soviet fast reactor, powered radar ocean reconnaissance satellites from 1970 to 1988 and scattered debris across northern Canada in January 1978. Sources 4, 10, 13, and 15.Krasnaya Zvezda built the reactor, a core of 37 uranium molybdenum fuel rods cooled by liquid sodium potassium alloy, with thermoelectric converters that delivered about 3 kilowatts of electricity. Sources 10, 11, and 15. The Soviet Union flew about 31 of these power plants at an altitude near 250 kilometers, where the radar worked, and boosted each reactor to about 900 kilometers at the end of the mission. Sources 1 and 10. The disposal maneuver failed on Cosmos 954 in 1978, Cosmos 1402 in 1983, and Cosmos 1900 in 1988. Sources 2, 10, and 17. The last launch flew on March 14, 1988. Sources 10 and 16.Classification: the BES 5 is treated as a power or demonstration reactor because its charter was to supply electric power to an operational radar satellite, so this report weights technical analysis above commercial analysis. Sources 4 and 16.Technically, the record shows a compact, shielded core that no one could repair, a coolant loop that had to be discarded at end of life, and a disposal system that worked on most flights and failed on three. Commercially, the record shows one documented bill: Canada spent about 14 million Canadian dollars, roughly $11.7 million in US Dollars, to search 124,000 square kilometers after a single reentry, and the Soviet Union paid about 3 million Canadian dollars, roughly $2.5 million. Source 13. The program left a modeled population of about 47,800 NaK droplets in orbit, and the Soviet records state no development or operating budget. Sources 4 and 6.Detailed FindingsTechnical AnalysisWhat Was the Design of the BES 5 Buk Reactor?The BES 5 was a fast neutron reactor with a compact core, a beryllium reflector, and a heavy shield, built to ride on a satellite of about 3,800 kilograms. Sources 10, 15, and 16. The core held 37 steel clad fuel rods and weighed 53 kilograms as an assembly. Sources 10 and 15. Reported core dimensions differ between sources, from about 0.20 meters across and 0.60 meters long to about 0.24 meters across and 0.67 meters long. Sources 10 and 15. A shield of lithium hydride and stainless steel, supplemented with tungsten and uranium alloys, brought the whole power plant to about 385 kilograms. Sources 10 and 15. The reactor ran at about 100 kilowatts thermal. Sources 10 and 15. Six movable control elements regulated the core, and sources describe them as beryllium cylinders and as boron carbide rods with lithium hydride inserts. Sources 10 and 15. A fast core needs no moderator, which removes the mass of graphite or water from a vehicle where every kilogram costs launch capacity. Source 15.Krasnaya Zvezda built the reactor. Source 11. The Leypunsky Institute for Physics and Power Engineering and Krasnaya Zvezda presented the flight record to the institute council in 2010, and that record dates the first flight design test to Cosmos 367, launched on October 3, 1970. Source 4.What Fuel, Coolant, and Power Conversion Did the Reactor Use?The fuel was a uranium molybdenum alloy enriched above 90 percent in uranium 235, and the coolant was a liquid sodium potassium alloy known as NaK. Sources 10 and 15. Sources differ on the fuel mass: about 30 kilograms in one account, 35 to 50 kilograms in another, and about 50 kilograms in a Soviet statement of 1978. Sources 10 and 15. A primary NaK loop held about 13 kilograms of alloy and a secondary loop held about 26 kilograms. Source 10.Semiconductor thermoelectric converters turned the heat into about 3 kilowatts of electricity, an efficiency of 3 percent, calculated from the reported 3 kilowatts electric and 100 kilowatts thermal. Sources 1 and 10. At 385 kilograms the plant delivered about 7.8 watts of electricity per kilogram, calculated from the same figures. Source 10. The remaining 97 kilowatts left the satellite as radiated heat, calculated from the thermal and electric ratings.What Engineering Problems Shaped the BES 5 Design?Radar physics forced the reactor into a low orbit, and the low orbit ruled out solar panels. Source 16. The satellites flew at 250 to 280 kilometers with a 65 degree inclination. Source 10. Solar panels would have lost altitude to drag and stopped producing power in the shadow of the Earth. Source 16. A reactor delivered power at every point in the orbit, at the price of a shield and an end of life problem. Sources 15 and 16.The end of life problem was that the low orbit decays within weeks after a satellite loses altitude control. The designers solved it by boosting the reactor to about 900 kilometers, a maneuver one contemporary account called fuel expensive. Sources 10, 12, and 15. The 1978 accident changed the design. From Cosmos 1176 in 1980, the fuel core separated from
- 4 days ago1h 3m
ASML EUV Lithography Systems
ASML EUV lithography systems are the only production tools printing leading edge chip features with 13.5 nanometer light, and ASML guided 2026 net sales near $50 billion. The company reported third quarter 2026 net sales of $13.0 billion, a gross margin of 57.5%, and EUV system sales of $4.6 billion, then guided full year net sales to between $49.1 billion and $50.2 billion. Sources 1 and 26. A low NA NXE:3800E scanner, where NA means numerical aperture, rates 195 to 220 wafers per hour at a price between $180 million and $300 million; a High NA EXE:5200B scanner resolves about 8 nanometers at a price near $380 million to $400 million. Sources 13, 15, 17 and 27.The light source sets the engineering agenda of 2026. ASML raised its tin plasma output from 600 watts to 1,000 watts in February 2026; ASML technologists project a path to 330 wafers per hour per tool by 2030 from about 220 today. Sources 18 and 19. High NA availability reached 84% in July 2026 against a 90% target for the fourth quarter; Intel reports more than 1 million wafers exposed on High NA tools. Sources 16 and 26. A scanner priced near $400 million is judged by how often it runs.The cost comparison that decides most High NA purchases comes from IBM: a single High NA exposure costs roughly 2.5 times a single low NA exposure, so High NA lowers cost only where it replaces three or more low NA masks. Source 14.ASML management stated on October 14, 2026 that planned 2027 and 2028 output will not fully meet customer demand, with low NA EUV and immersion capacity rising 30% in 2027. China is guided at about 20% of 2026 net sales, down from 33% in 2025. Sources 17 and 26. Buyers and shareholders alike are reading a supplier whose order book runs ahead of its shipping schedule.What Is EUV Lithography and Who Sells It?EUV lithography is a projection printing process that uses 13.5 nanometer light from a laser produced tin plasma to expose patterns onto photoresist coated silicon wafers inside a vacuum tool. Sources 6 and 27. EUV names the band of light near 13.5 nanometers, far shorter than visible light or the 193 nanometer light of earlier scanners. The wavelength is 14 times shorter than the 193 nanometer light of immersion deep ultraviolet scanners, a ratio calculated from the two wavelengths, and photons at 13.5 nanometers carry about 92 electron volts of energy. Source 27.The engineering classification is optical projection lithography built on three domains of physics: plasma physics for the source, multilayer thin film optics for the reflective imaging train, and photochemistry for the resist. No material transmits 13.5 nanometer light in useful amounts. Every optical element is therefore a mirror, the mask is reflective, and the entire light path runs in vacuum with a low pressure hydrogen atmosphere around the source collector. Sources 8 and 27.ASML is the only company that sells production EUV scanners. It started in 1984 as a joint venture of Philips and Advanced Semiconductor Materials International, shipped its first prototype EUV tool, the NXE:3100, in 2010, and shipped its 100th EUV system in 2020. Source 11. Competing source and tool concepts, including a free electron laser effort in the United States and a government backed prototype program in China, remain at the prototype or financing stage. Sources 28 and 29. The market has one seller.Which Scientific Principles Set the Limits of EUV Lithography?Resolution in projection lithography follows the Rayleigh relation: critical dimension equals k1 times wavelength divided by numerical aperture, and EUV improves it by shortening the wavelength and by raising the aperture. With a process factor k1 near 0.32, the calculation gives 13.1 nanometers at 0.33 NA and 7.9 nanometers at 0.55 NA, which matches the 13 nanometer and 8 nanometer figures published for the two platforms. Sources 13 and 16. Depth of focus falls with the square of the numerical aperture, so High NA forces thinner resist films. Source 14.Two physical facts fix the 13.5 nanometer wavelength. Tin plasma emits strongly near that wavelength, and molybdenum and silicon multilayers reflect it. A molybdenum silicon mirror uses 40 to 50 layer pairs of about 3.5 nanometers each and reflects about two thirds of the incident light at its peak, with a theoretical limit near 75%. Sources 24 and 27. An optical train of about nine reflections passes a small fraction of the source light: 70% reflectivity per surface leaves 4.0% after nine reflections, a calculation that agrees with published analyses putting total mirror absorption near 96%. Source 27. Most of the light never reaches the wafer. Conversion of laser energy into in band EUV light reaches only about 5% to 6%. Source 27.Photon statistics impose the hardest limit. At ASML’s rated dose of 30 millijoules per square centimeter, about 20 photons strike each square nanometer, a figure calculated from
- 5 days ago26 min
AI Model Retirement Migration
An AI model retirement migration before OpenAI’s October 23, 2026 shutdown of GPT 4 and o1 requires frozen test cases, repeated trials, and regression gates set per item.OpenAI announced the shutdown on April 22, 2026. The list covers GPT 4, GPT 4 Turbo, o1, o1 pro, o3 mini, o4 mini, GPT 4.1 nano, GPT 3.5 Turbo, a May 2024 snapshot of GPT 4o, and every fine tuned snapshot built on them, with GPT 5.6 Sol, Terra, and Luna named as successors. Google stopped serving Gemini 2.5 Pro, Flash, and Flash Lite on its Gemini Enterprise Agent Platform on October 20, 2026, and named Gemini 3.8 Flash as the replacement.A 2026 study of 900 benchmark items, each run 50 times across three consecutive GPT releases, found reliable improvements and reliable regressions inside every migration. In one comparison the aggregate score rose 7.3 points and 6.0 percent of items regressed anyway. A single run per item detected none of the 457 confirmed changes.Price adds a second variable. Moving a GPT 4 workload to Sol cuts list input cost 83.3 percent. A GPT 3.5 Turbo workload that lands on Terra pays eight times the old output rate, and Luna, at $0.20 input and $1.20 output per million tokens, sits below the old GPT 3.5 Turbo rate.Teams should freeze their own test cases from production traces, run 30 to 50 trials on high consequence items, gate cutover on the list of regressed items, and schedule the next wave: December 11, 2026 and January 14, 2027.Definition, Scope, and Technical ContextWhat Is an AI Model Retirement Migration?An AI model retirement migration is the engineering process of moving a production workload from a hosted model identifier that a provider has scheduled to stop serving to a successor identifier, while holding behavior, cost, and latency inside tolerances the team defines in advance. It belongs to software change management, and the dependency being changed is a probabilistic function whose internals the deployer cannot inspect. Deprecation is the announcement, retirement is the date the identifier stops answering requests, and drift is a behavior change that arrives with no announcement at all.The October 23, 2026 cutoff comes from OpenAI’s April 22, 2026 notice. The identifiers include GPT 4 (the 0613 snapshot), GPT 4 Turbo (the April 9, 2024 snapshot), a GPT 4 1106 preview, a GPT 4o snapshot dated May 13, 2024, GPT 4.1 nano, o1, o1 pro, o3 mini, o4 mini, GPT 3.5 Turbo, the completions variants of several of these, GPT Image 1, and the fine tuned versions built on GPT 3.5 Turbo, GPT 4, GPT 4.1 nano, and o4 mini. OpenAI named GPT 5.6 Sol as the successor for GPT 4, GPT 4 Turbo, o1, o3 mini, and the older GPT 4o snapshot, GPT 5.6 Terra for o4 mini and GPT 3.5 Turbo, GPT 5.6 Luna for GPT 4.1 nano, and a GPT Image 2.5 model for GPT Image 1.Today is October 22, 2026, so a team reading this has one day left. Calls to a retired identifier stop being served after the cutoff, which converts every unfinished migration into a production incident. The work described below fits a one day sprint for a small inventory and a multi week program for a large one, and the same pipeline serves the later waves.Which Providers Retire Models Between October 19 and October 23, 2026?Four providers retire models inside that window. Google ended calls to Gemini 2.5 Pro, Gemini 2.5 Flash, and Gemini 2.5 Flash Lite on its Gemini Enterprise Agent Platform, formerly Vertex AI, on October 20, 2026, while the Gemini API keeps serving those models with no shutdown date. Google shuts down Gemini Omni Flash Preview and three Veo 3.1 preview models on October 22, 2026, with Gemini Omni 1.1 Flash as the replacement, and retires a deep research preview agent on October 23, 2026. Together AI ends serverless access to Llama 3.3 70B Instruct Turbo and DeepSeek V4 Flash 0731 on October 22, 2026. NVIDIA stopped supporting Nemotron 3 Super on its free NIM endpoints after October 19, 2026.How Much Notice Do Providers Give Before Retiring a Model?OpenAI publishes a minimum of six months of notice for generally available models, three months for specialized variants such as chat, Codex, and deep research models, and as little as two weeks for preview models. The April 22, 2026 announcement preceded the October 23, 2026 shutdown by six months and one day. Fine tuned models follow the schedule of their base model, and the List models and Retrieve a model endpoints return a shutdown_date field that a script can read.Google labels the dates on its Gemini API deprecation page as the earliest possible shutdown dates and sets no minimum notice period. Anthropic commits to preserve the weights of every publicly released Claude model for at least the lifetime of the company, and the same commitment page states that preserved weights do not guarantee continued availability for inference. The analysis here covers hosted model APIs and excludes consumer chat product
- 6 days ago27 min
Biotech Critical Infrastructure Bills
The biotechnology and biological data critical infrastructure bills introduced September 24, 2026 direct CISA to support biomanufacturers and protect genomic data systems, and they impose no duties on companies.Each bill amends section 2202 of the Homeland Security Act of 2002, codified at 6 U.S.C. 652, which defines what CISA does. The first, H.R. 10569 and S. 5502, adds biotechnology and biomanufacturing infrastructure to CISA’s outreach, risk assessment, and national planning duties without creating a new critical infrastructure sector. The second, H.R. 10568 and S. 5522, tells CISA to treat genetic data systems that involve genomic sequences and other sensitive biometric data as critical infrastructure, to consult industry on security protocols, and to staff the work. Both track recommendation 2.4a of the National Security Commission on Emerging Biotechnology, which sunsets in December 2026.For a startup that runs fermentation or cell culture production, or that stores sequence data, the practical change is a federal counterpart at CISA. No rulebook comes with it. The binding obligations arrive from other sources: the Justice Department bulk data rule at 28 C.F.R. part 202, the BIOSECURE provisions in section 851 of the FY2026 National Defense Authorization Act, the military contractor cybersecurity certification rule, and the CIRCIA final rule that CISA sent to the White House budget office on October 1, 2026. The bills matter because they show customers, investors, and insurers where the federal government is looking, and diligence questions follow that attention.The Senate text reviewed for this report is a draft with sections missing, and it does not define genetic data systems or biotechnology infrastructure. Those definitions are the first items to watch. Two related amendments to the FY27 defense authorization bill were filed in June 2026, and that bill is the likeliest vehicle before the commission ends. For founders, general counsel, and investors, the sensible posture in October 2026 is to build the security and data controls that existing rules already demand, and to treat any CISA outreach as a free channel worth joining.Detailed FindingsWhat Do the Two Bills Actually Require?Each bill amends the Homeland Security Act of 2002 and requires CISA to act, while leaving private companies with no new obligation. The biotechnology and biomanufacturing bill adds owners and operators of that infrastructure to the entities CISA must coordinate with under subsection c, paragraph 2 of 6 U.S.C. 652. Where the infrastructure has no designated sector risk management agency, CISA must assign staff to provide analysis, expertise, and technical assistance, to conduct outreach, and to convene the relevant agencies. A new paragraph 14 directs CISA to take the actions necessary to protect it.The national risk assessments that feed the plan under section 2202, subsection e, paragraph 1 would be carried out in coordination with existing information sharing and analysis organizations, and the text names BIO ISAC and H ISAC along with industry associations. The bill also adds biotechnology and biomanufacturing infrastructure to the list of sectors the national plan covers. The Senate draft sets a work plan due 45 days after enactment, an implementation report one year later, and a report on emerging biotechnology that must say whether any critical biotechnology areas remain unprotected.That draft work plan also calls for an update to the national strategic plans by September 30, 2026. The date falls six days after introduction and before the 45 day clock could start, so the final text has to change it. The mismatch is one reason this report treats the reviewed language as a draft.Do the Bills Create a New Critical Infrastructure Sector?No. The commission asked Congress to have DHS cover biotechnology under the existing sectors and to say so explicitly if some critical biotechnology areas fit none of them, and press accounts of the September 2026 bill describe plans without a new sector. Sixteen sectors exist. PPD 21 set them in 2013, and NSM 22 of April 30, 2024 kept them without addition, although a 2022 DHS report had raised Space and Bioeconomy sectors.The sector label matters because biotechnology touches health, agriculture, and industry at once. A plant that ferments a food ingredient may sit under food and agriculture, one that makes a specialty chemical under chemical, one that builds equipment under critical manufacturing, and one that supplies a drug precursor under healthcare and public health. The label selects the sector risk management agency, the reporting criteria, and the regulator that eventually asks questions.How Would the Biological Data Bill Treat Genomic Data Systems?The biological data bill makes CISA responsible for the security of genetic data systems that involve genomic sequences and other sensit
Get The Innovation Attorney Podcast summaries in your inbox
Free AI-powered daily recaps. Key takeaways, quotes, and mentions — in a 5-minute read.
Get Free Summaries →Free forever for up to 3 podcasts. No credit card required.
You Might Also Like
Listeners also like.
AI Sidebar
Experts discuss AI’s impact on law, business, and society, exploring accountability, regulation, and the future of legal practice.
Scaling Laws
Experts analyze AI policy, law, and governance through interviews and breaking news analysis.
The AI in Business Podcast
Interviews with AI executives reveal practical applications and strategic insights for business leaders adopting AI.
The AI XR Podcast
Insightful interviews with tech founders and executives on AI, virtual reality, augmented reality, and the future of spatial computing.
No Priors: Artificial Intelligence | Technology | Startups
Discusses AI advancements and implications with researchers, engineers, and founders shaping the future of artificial intelligence.
TechStuff
TechStuff explores how technology, AI, and the internet reshape culture, identity, and power, from Silicon Valley to everyday life.
Possible with Reid Hoffman
Explores optimistic, realistic futures through in-depth conversations with innovators in AI, science, health, and global affairs.
This Week in AI
Four AI experts analyze the latest news and trends in artificial intelligence each week.
TED Tech
Experts examine how technology reshapes society, science, design, and business through TED Talks hosted by Sherrell Dorsey.
Unriveted
Discusses technology, AI, machine learning, data engineering, cloud services, and the startup experience.
80,000 Hours Podcast
Discusses artificial intelligence and global catastrophic risks with experts and researchers.
"The Cognitive Revolution"
Explores the transformative impact of artificial intelligence through interviews with innovators shaping its future.
About The Innovation Attorney Podcast
The Innovation Attorney is a publication about the intersection of venture capital, technology, public policy and law. theinnovationattorney.substack.com
Customized Recaps
AI-powered recaps with compact key takeaways, quotes, and insights.
Straight to Your Inbox
Get key takeaways from The Innovation Attorney Podcast in a 5-minute read.
Save Hours Every Week
Stay current on your favorite podcasts without falling behind.
Frequently Asked Questions
What is Podzilla's The Innovation Attorney Podcast daily summary?
It's a free AI-powered email that summarizes new episodes of The Innovation Attorney Podcast as soon as they're published. You get the key takeaways, notable quotes, and links & mentions — all in a quick read.
How does the The Innovation Attorney Podcast podcast summary work?
When a new episode drops, our AI transcribes and analyzes it, then generates a personalized summary tailored to your interests and profession. It's delivered to your inbox every morning.
Is this an official The Innovation Attorney Podcast product?
No. Podzilla is an independent service that summarizes publicly available podcast content. We're not affiliated with or endorsed by The Innovation Attorney.
Can I get summaries of other podcasts too?
Absolutely! The free plan covers up to 3 podcasts, any in the catalog. Basic raises that to 15, Pro to 40 and Max to 100. Browse our full catalog at /podcasts.
What topics does The Innovation Attorney Podcast cover?
The Innovation Attorney Podcast covers topics including Technology, Business, Investing. Our AI identifies the specific themes in each episode and highlights what matters most to you.
Start getting The Innovation Attorney Podcast summaries tomorrow morning.
Free forever for up to 3 podcasts. No credit card required.
Free forever for up to 3 podcasts. No credit card required.