In partnership with
Nobody's coming to approve your idea
People assume that real businesses get built by other people with more resources. It's how great ideas get stuck in notes apps and conversations.
But you don't need six months and a developer's salary to validate your concept. You just need a weekend. Describe what you want to exist and Lovable builds it: a real site, a real product, ready to accept real payments from real customers. Not a mockup to show people. An actual thing they can use.
You don't need anyone's sign-off to get started. Take $5 off your first month.
Discount of $5 applies to first month of any paid plan. You need to sign up via this webpage in order to claim the discount. The discount applies to new and returning users without an active paid plan. Terms apply.
⚖️ POLICY & REGULATION
Warren Gives Anthropic Until Nov. 3 To Answer Four Pages Of Questions On The Catastrophic Risks Its S-1 Discloses: Sen. Elizabeth Warren (D-Mass.) has asked CEO Dario Amodei for "exhaustive responses" by Nov. 3 to questions on the filing's warning of "catastrophic or existential risk to humanity," on Anthropic's relationships with Amazon and Broadcom, and on how it would justify a valuation she frames as approaching $2 trillion, the Wall Street Journal reported Wednesday, Oct. 28, after viewing the letter. The Journal reports the company is targeting the listing for the week of Nov. 16. The letter stacks on the FTC's demand for executive testimony, which we carried Tuesday, and NYC Comptroller Mark Levine's Oct. 14 letter urging a pause, over which Anthropic has since opened talks with his office. No release had posted on Warren's site by this morning. Link
The U.S. AI Testing Center Runs On About $10 Million To $15 Million A Year As Britain Adds £120 Million To Its Institute On Budget Day: Semafor: Semafor reported Wednesday that the U.S. counterpart at Commerce gets roughly $10 million to $15 million a year, that the administration's FY2027 request is $27 million, and that a Trahan-Obernolte bill would authorize $100 million a year through 2029; its director, Chris Fall, resigned in July, while the U.K. AI Security Institute has more than 100 staff. The comparison lands as Britain presents today's Budget, which carries the £621 million AI security package we led with Monday, including an extra £120 million for the institute. The gap is the argument for the Trahan-Obernolte authorization when Congress returns Nov. 9. Link
Australia's House Passes A Digital Duty Of Care Bill That Reaches AI Chatbots And Bars Them From Standing In For Children's Relationships: The Online Safety Amendment (Digital Duty of Care) Bill 2026, introduced Oct. 15 by Communications Minister Anika Wells, passed the House of Representatives Wednesday, Oct. 28, AAP reported. It puts one statutory duty on social media, games and AI chatbot services to keep users safe "so far as is reasonably practicable," with maximum penalties above A$100 million, gives users over 16 an algorithm opt-out, and replaces the Basic Online Safety Expectations and industry codes after a 12-month transition. An amendment from independent MP Kate Chaney adds "attachment hacking" provisions requiring platforms to keep chatbots from replacing children's human relationships. A Senate inquiry hearing is set for Nov. 9 with a report due late November; Labor needs the Greens, who want higher penalties and a right to compensation, to pass it before Christmas. Link
STATES
Hochul Signs New York's Ban On Unsafe AI Companion Features For Minors, With AG Penalties Up To $25,000 Per Violation From Jan. 1: Gov. Kathy Hochul signed S 9051-B (Sen. Kristen Gonzalez, Assembly Member Alex Bores) on Tuesday, Oct. 27, Fortune reported. From Jan. 1, an operator of any generative AI system that gives "ongoing, adaptive responses" may not offer a minor twelve listed features, among them claiming to be human or to have emotions, claiming a romantic or authority relationship, sycophancy, unprompted emotional questions, recalling personal or health details from earlier sessions, encouraging secrecy from parents, promoting self-harm and sexual content, unless an age check shows the user is an adult. Operators must offer more than one age-assurance method, at least one that does not rely on government ID, and delete the data after use; the AG can seek $25,000 per violation and order unlawfully collected data, and models trained on it, destroyed. "No company should be peddling chatbots to kids encouraging self-harm or pretending to be their friend or more," Hochul said. The Docket, below, has the coverage test and the duties. Link
Pennsylvania Senate Bill Would Bar Schools From Using AI To Deliver Student Mental Health Services: Sen. Nick Miller's SB 1505, referred to Senate Education on Monday, Oct. 26, would prohibit schools from using AI to provide mental or behavioral health services otherwise performed by licensed professionals, while letting school mental health staff use it for administrative work, WTAJ reported. It is the Senate companion to Rep. Melissa Shusterman's HB 2787, which would also bar AI vendors from contracting with schools for such systems; trackers disagree on where that bill sits. Colorado, Maine, Rhode Island, Tennessee and Vermont restricted AI-delivered therapy this year. Link
New Hampshire Lawmakers Request A Batch Of Data Center Bills For 2027, As Chatham County Residents Open A Push For A Two-Year North Carolina Pause: At least three incumbent New Hampshire legislators have asked the Office of Legislative Services to draft data center bills for the session that opens in January, including a pause sought by Sen. Daniel Innis (R-Bradford), New Hampshire Bulletin reported Tuesday; Gov. Kelly Ayotte said in September she will put a likely multi-year moratorium in her budget. No bill numbers are assigned yet. In North Carolina, residents and the Haw River Assembly hold a press conference in Pittsboro at 2 p.m. today calling for a two-year statewide moratorium, the first of several such events planned across the state, WRAL reported; Chatham County's own pause runs to Feb. 11, 2027. Link
Nine Local Decisions: San Diego And Pittsburg Pause For 45 Days, Venice And Martin County For A Year, Pulaski County Overrides A Veto, And Iron County Ends Utah's First Moratorium: San Diego's council voted unanimously Tuesday, Oct. 27, to ratify Mayor Todd Gloria's 45-day moratorium on data center permits, approvals and development, extendable to nearly two years, and gave staff 180 days to bring rules; it is largely symbolic for the 400 MW NAVWAR proposal, which sits on federal land. Pittsburg, Calif., adopted its own 45-day pause unanimously Monday, exempting the first phase of Avaio's Perseus project. In Florida, Venice adopted a 12-month moratorium on projects with 50 MW or more of peak load and Martin County a one-year pause on large data centers in unincorporated areas, though a hyperscale application filed Oct. 12 is not affected. Mountain Home, Idaho, adopted an emergency moratorium unanimously after a developer floated about 3.4 million square feet at Bennett Ranch; Pulaski County, Ark., overrode County Judge Barry Hyde's veto of its disclosure ordinance 10-4 as AVAIO threatens suit over its $6 billion Wrightsville project; and Iron County, Utah, repealed its moratorium early and adopted an overlay requiring a development agreement above 50 MW. Mexico, Mo., extended its pause to Nov. 30 ahead of a Nov. 23 vote, and Clinton, Iowa, passed the first of three readings on rules for QTS's roughly 1,100-acre campus. Link
FROM THE DISPATCH TEAM
Join Docket Room, free: every bill in America inside Claude or ChatGPT
Chatbot laws like New York's (States, above) travel in packs. The map shows 21 bills in 16 states and Congress that share the text of Washington's and Oregon's chatbot laws: five already law, three through one chamber, eleven waiting to come back. We pulled it from Docket Room's MCP this morning with one question. Add it to Claude, ChatGPT, Cursor or VS Code and ask in plain English: which states copied this bill, what moved overnight on the bills I follow, what hearings are next week. All 50 states, DC and Congress, read nightly from official sources.
Free accounts read everything and track 5 bills. MCP server: docketroom.ai/api/mcp, OAuth sign-in, nothing to install.
🛡️ SAFETY & SECURITY
Anthropic's S-1 Warns A Catastrophic AI Incident Could Exceed What The Insurance Market Can Cover, And Discloses Uncapped Indemnities To Outside Evaluators: Among the risk factors in the filing made public Oct. 26, Anthropic warns that losses from a catastrophic AI event could surpass what any single insurer, and possibly the whole market, can underwrite, describes the market for AI-liability cover as nascent, says insurance did not fully cover its $1.5 billion Bartz copyright settlement, and discloses that several agreements require it to indemnify outside evaluators against certain third-party claims with no cap, Insurance Business reported Tuesday. "It cannot be assumed that all AI exposures will be insurable," WTW's Andrew Hill told the outlet. The disclosure feeds directly into the AI product-liability bills states are expected to file in 2027 and into Warren's Nov. 3 questions (Policy, above). Link
Researchers Propose A Scientist-Led International Network To Draft Frontier AI Safety Standards: A paper published Tuesday, Oct. 27, in Frontiers in Artificial Intelligence by Daniele Gerundino, Yishu Mao, Leonardo Chiariglione and Zheng Liang proposes a network of scientists to complement intergovernmental efforts: it would coordinate safety research, build audit, benchmark, monitoring and hardware-enabled governance tools, and draft standards and conformity assessments for formal standards bodies to adopt. The authors call it "a call to action rather than a detailed project plan." Link
📜 DISPATCH PRO - enjoy for free today!
Open to every reader today. This is the block Dispatch Pro subscribers get every morning.
🌙 Since Tuesday's issue
🔴 Act · 18. NY S 9051-B, unsafe AI companion features for minors: signed by Gov. Hochul Tuesday, Oct. 27 (Fortune); takes effect Jan. 1, 2027. Stage 8, coverage 3, copies 5, penalty 2. The Bill, below.
⚪ Noise · 1. PA SB 1505, Sen. Nick Miller's ban on schools using AI to deliver mental and behavioral health services: referred to Senate Education Monday, Oct. 26; not yet in Docket Room's record. Companion to Rep. Shusterman's HB 2787. Stage 1.
Docket Room's nightly check found no status change, hearing notice or amendment on the AI bills we track; the New York signing below came from the governor's Tuesday announcement as reported by Fortune, and Docket Room's record still reads passed one chamber. Act is 12 or more, Watch is 7 to 11, Noise is under 7. The score adds up stage, coverage, copies in other states, a date inside ten days, momentum and penalties, and we print it so you can disagree with it.
3️⃣ The Three
NY S 9051-B, unsafe AI companion features for minors: signed Oct. 27, in force Jan. 1, 2027, AG penalties up to $25,000 per violation. Act, 18. Today's Bill.
CO SB 26-189, the Automated Decision-Making Technology Act (Ch. 131): takes effect Jan. 1, 2027 if the Attorney General finishes its rules; the hearing and comment deadline were Oct. 26 and final rules are next. Watch, 11.
NJ A 3966, DEP study of data center water use: on the Assembly's voting-session agenda Thursday, Oct. 29. Watch, 10. Monday's Bill.
📜 The Bill · NY S 9051-B, Prohibition on unsafe AI companion features for minors · 🔴 Act 18
Coverage test. You are covered if you are an operator that provides a covered AI companion to a user in New York, defined as "a generative artificial intelligence system with a natural language interface, including via writing or sound, that provides ongoing, adaptive responses to user inputs" (Sec. 1800(5)). That reaches general-purpose assistants, not only companion apps, and "operator" includes affiliates, subsidiaries and beneficial owners (Sec. 1800(6)). You are out of scope for a bot used solely for customer service that cannot answer off-topic, solely for efficiency or technical assistance that cannot answer off-topic, or for internal or employee-productivity use (Sec. 1801(2)). Conduct is out of state only if the user is physically outside New York (Sec. 1805).
Obligations
Do not provide any of the twelve listed unsafe AI companion features to a user unless you have determined, by a permitted age assurance method, that the user is not a minor you know to be under 18. The list includes claiming to be human or to feel emotions, claiming a personal, professional or authority relationship, flattery or sycophancy, unprompted emotional questions, reusing health or personal information more than twelve hours old or from an earlier session, encouraging secrecy or isolation, engagement optimization that overrides safety guardrails, self-harm or eating-disorder promotion, and sexually explicit output (Secs. 1800(8), 1801(1)). Due Jan. 1, 2027. Penalty: AG injunction, restitution, disgorgement including destruction of unlawfully obtained data and any algorithm trained on it, damages, and civil penalties up to $25,000 per violation (Sec. 1802(1)). Not a verified quote; treat as our reading.
Offer more than one age assurance method, including at least one that does not rely on government ID or that lets the user stay anonymous to you; self-declaration counts only when a user declares they are a minor (Sec. 1804(1)). Due Jan. 1, 2027, unless AG rules under Sec. 1803 or Article 45 take effect first. Penalty: Same as above (Sec. 1802(1)). Not a verified quote; treat as our reading.
Use age assurance data only to determine minor status and delete it immediately after the attempt, except where other law requires keeping it (Sec. 1804(2)). Due Jan. 1, 2027. Penalty: Same as above (Sec. 1802(1)). Not a verified quote; treat as our reading.
Do not rely on terms of service or adhesion contracts to waive or shift liability under the article; such clauses are void, and affiliates that structured themselves to avoid liability can be held jointly and severally liable (Sec. 1802(3)-(4)). Due Jan. 1, 2027. Penalty: Same as above (Sec. 1802(1)). Not a verified quote; treat as our reading.
Delta. New York has no rule today on what a chatbot may say to a minor; the 2025 budget's AI companion law (GBL Art. 47) requires suicide-ideation protocols and periodic notices that the user is talking to AI, and Sec. 1804(3) keeps it in force alongside the new article. The nearest enacted cousin, Georgia SB 540 (similarity 0.801), is a disclosure law; Washington's HB 2225 (0.769) adds crisis protocols and recurring disclosures. New York goes further than either: it bars the companion features themselves for known minors and ties every adult experience to an age check.
Owners. Trust and safety and model policy (the twelve features are output behaviors, so they live in system prompts, classifiers and fine-tuning), product and growth (engagement optimization and memory across sessions), identity and age assurance, privacy (deletion of age data), and legal for the customer-service and internal-use carve-outs.
Similar bills.
GA SB 540, conversational AI disclosures: signed (similarity 0.801).
WA HB 2225, AI companion chatbots: enacted, effective Jan. 1, 2027 (0.769).
HI SB 3001, AI companions and minors: enacted as Act 248 (0.768).
PA SB 1090, AI disclosures and safeguards: passed the Senate (0.773).
US HR 8623, GUARD Act: in committee (0.781).
Next date. Friday, Jan. 1, 2027: the article takes effect, the same day as Washington's and Colorado's chatbot laws. Until the Attorney General adopts age assurance rules, operators must rely on their own "reasonable" method meeting Sec. 1804's tests, so the design choices made this quarter are the ones that will be judged.
Verification. Docket Room rates the record single-source and still shows it as passed one chamber (amended to 9051-A May 11); the signing on Oct. 27 is reported by Fortune, which quotes the governor, and the Senate's bill page has not yet posted it. Every obligation above is our reading of the ingested text, section by section; Docket Room's quote-verified impact assessment is a Team-plan tool not available on this connection, so 0 of 4 rows are tool-verified. We could not confirm whether a chapter amendment was agreed at signing. Score: stage 8 (effective date set), coverage 3 (any generative AI system with ongoing adaptive responses), copies 5 (MO, GA, LA, HI, FL and more above 0.75, capped), dated event inside ten days 0, momentum 0 (fast_mover false), penalty 2 ($25,000 per violation). No bill with a dated event inside ten days qualified: NJ A 3966 (Thursday's floor vote) was Monday's Bill, so the highest-scoring bill on the ladder is featured instead.
For your team. New York's governor signed S 9051-B on Oct. 27, and from Jan. 1 any generative AI chat product reachable in New York may not show minors twelve companion behaviors, from claiming feelings to cross-session memory of personal details, unless we have age-checked the user as an adult. The AG can seek up to $25,000 per violation and order data and models trained on it destroyed, and customer-service and internal bots are carved out only if they cannot answer off-topic. We already run age signals and some of these guardrails for Washington's Jan. 1 law; we need a New York gap list and an age assurance method that does not depend on government ID before year end.
📅 Dates
This week's AI hearings
Everything this week is on the board below, led by Thursday's New Jersey Assembly voting session with A 3966 on the agenda; Docket Room's calendar shows no other AI bill hearing in any state before Nov. 7.
Deadline Board · next 30 days
Oct. 29: PJM deadline to file its revised reliability backstop procurement proposal at FERC, after DOE's Oct. 7 statement of position (US)
Oct. 29: New Jersey Assembly voting session; A 3966 (DEP study of data center water use) is on the agenda for a floor vote after Appropriations reported it Oct. 22 (NJ) · Docket Room
Oct. 29: San Antonio City Council final vote on data center zoning amendments (specific-use authorization, 500-foot buffers, Edwards Aquifer recharge zone ban) in place of a moratorium (TX)
Oct. 30: Comments due to Rep. Dingell on her federal Data Center Bill of Rights framework (US)
Oct. 31: PJM stakeholder review of ride-through rules for 50-MW loads runs through October; FERC filing targeted for November (US)
Oct. 31: Meta's answers due to Sen. Wyden and six Senate Finance Democrats on R&D tax credits claimed for AI data centers described as "pilot models" (US)
Oct. 31: Amazon withdraws its Bridgefield Tech Center application for the former GW campus in Ashburn by month-end under Loudoun's Project Potomac Promise deal (announced Oct. 20) (VA)
Nov. 1: PJM: facilities entering the study queue must show ride-through compliance (US)
Nov. 2: Westfield, Mass.: Servistar's 20-day window to appeal the Planning Board's Oct. 13 denial of its data center permit extension closes (MA)
Nov. 3: Election Day: 41 local data center measures in 10 states (Ballotpedia count as of Oct. 21), 18 to 20 in Ohio (US)
Nov. 3: Littleton, Colo., public hearing and second reading on the six-month data center moratorium (CO)
Nov. 3: Ottawa, Kan., voters decide the 10-year ban on data centers over 25 MW, left on the ballot by Judge Godderz's Oct. 20 ruling (KS)
Nov. 3: Florida PSC decision on Duke Energy Florida's data center rate proposal, which commission staff recommended denying Oct. 22 (document 07479-2026) (FL)
Nov. 3: Anthropic's answers due to Sen. Warren on four pages of questions about the S-1's catastrophic-risk disclosures, its Amazon and Broadcom ties and its valuation (per WSJ) (US)
Nov. 4: Michigan City, Ind., council second reading of the data center moratorium running to July 31, 2028, with Google's Royal Road campus exempt; third reading and final vote to follow (IN)
Nov. 5: Guilford County, N.C., commissioners resume the 180-day data center and crypto-mining moratorium (to April 12, 2027) after postponing the Oct. 15 vote (NC)
Nov. 9: Senate returns after the midterms; permitting deal (Lee, Capito, Heinrich and Whitehouse) with data center transmission cost rules awaits a vote (US)
Nov. 9: SEC roundtable on machine-understandable data and "super intelligence," 9 a.m. to 3:30 p.m., 100 F Street NE and livestreamed (Release 2026-106) (US)
Nov. 9: D.C. Council Judiciary and Public Safety hearing on B 26-0524, the Distribution of False Sexual Imagery Prohibition Amendment Act, 10:30 a.m. (DC) · Docket Room
Nov. 9: Australian Senate inquiry into the Online Safety Amendment (Digital Duty of Care) Bill 2026, which passed the House Oct. 28 and covers AI chatbots; report due late November (AU)
Nov. 9: Mexico, Mo., public hearing and 4 p.m. comment deadline on its data center ordinance (MO)
Nov. 12: Gilbert, Ariz., council revisits the data center zoning amendment (650-foot residential buffer, no drinking water for cooling) postponed Oct. 20 (AZ)
Nov. 13: Oregon PUC vote on Pacific Power's data center cost-assignment settlement under HB 3546 (OR) · Docket Room
Nov. 14: Dixon, Calif., 45-day data center and battery storage moratorium expires unless extended (CA)
Nov. 15: FERC: RTO and ISO responses due on pro forma large-load cost-recovery agreements (mid-November) (US)
Nov. 15: N.C. Utilities Commission decision expected on the Duke Energy large-load tariff settlement with the Public Staff, Amazon, Google, Meta and Microsoft (mid-November) (NC)
Nov. 16: Newsom's deadline for agency recommendations on lab monitors and an AI kill switch (CA)
Nov. 16: Raleigh staff publish draft data center regulations for public feedback (NC)
Nov. 16: Nevada PUC Phase 3 hearings on NV Energy's 20-year resource plan; final order targeted for mid-December (Docket 26-05007) (NV)
Nov. 16: Pitt County, N.C., commissioners' public hearing on a temporary data center moratorium (set unanimously Oct. 19) (NC)
Nov. 16: Week Anthropic is targeting for its IPO (per WSJ) (US)
Nov. 17: Loudoun County Board of Supervisors vote on the resolution pausing legislative data center and substation applications for up to 12 months, retroactive to Sept. 15 (deferred from Oct. 20 for more staff work) (VA)
Nov. 18: Memphis City Council budget session with MLGW on the three-year water and gas increases, including the new data center water class at 30 percent a year; committee Nov. 24; rates effective January 2027 if approved (TN)
Nov. 19: EPA's written and oral response due to Sen. Warnock and Rep. Hank Johnson on regulating diesel backup generators at data centers (letter dated Oct. 22) (US)
Nov. 19: New York City DCWP public hearing, 11 a.m., and written comment deadline on its proposed automated customer service rule (AI disclosure, one-step transfer to a human, three-year records) (NY)
Nov. 20: San Francisco and Oakland 45-day data center moratoriums expire unless extended (CA)
Nov. 20: San Joaquin County, Calif., 45-day data center moratorium (unincorporated areas) expires unless extended (CA)
Nov. 23: Mexico, Mo., council vote on its data center ordinance, 6 p.m. (MO)
74 more dated items beyond 30 days are on the full board. New dates are added from every issue; passed dates roll off.
🏗️ Local Moratorium Tracker · changed this week
San Diego, CA: moratorium; 45 days; extendable to nearly two years; data center permitting, approval and development citywide; adopted unanimously Oct. 27, ratifying Mayor Gloria's executive order; staff has 180 days to bring rules.
Venice, FL: moratorium; 12 months or until code changes; 50 MW monthly peak load at one site; adopted on final reading Oct. 27.
Martin County, FL: moratorium; 12 months; large data centers (50 MW peak), unincorporated areas; Oct. 12 hyperscale application not affected; adopted on second reading Oct. 27.
Mountain Home, ID: emergency moratorium; up to 182 days (state cap); data center applications and permits; adopted unanimously Oct. 27.
Clinton, IA: zoning ordinance (pending); data centers; QTS ~1,100-acre campus; first of three readings approved Oct. 27.
Alachua, FL: moratorium; 12 months; AI data center applications; adopted 5-0 on second reading (WCJB, Oct. 27).
Pulaski County, AR: regulations (noise, fire, water); high-intensity digital infrastructure; adopted 11-4 Sept. 22; vetoed by County Judge Hyde; veto overridden 10-4 Oct. 27.
Pittsburg, CA: moratorium; 45 days; extensions to two years need a 4/5 vote; new data centers; Avaio Perseus Phase I exempt; adopted unanimously Oct. 26, effective immediately.
Iron County, UT: overlay zone (moratorium repealed); 50 MW IT load needs review and a development agreement; water plan above 50,000 gal/day; moratorium repealed and overlay adopted unanimously Oct. 26 (moratorium had run to Nov. 22).
Mexico, MO: moratorium (extended); to Nov. 30; data centers; extended Oct. 26; ordinance vote Nov. 23.
Van Buren, AR: zoning rules; Industrial-2 only; 1,000-foot setback from residential districts; 65 dBA day, 55 dBA night; ordinance adopted 6-0.
Aurora, CO: zoning restriction (mixed-use office ban); application moratorium (proposed); permanent; data centers removed as a permitted use in the MU-OI district; still by right in airport, business/tech and medium industrial districts; MU-OI removal adopted unanimously Oct. 12; indefinite application moratorium failed Oct. 26, 6-5 short of the two-thirds needed to revive a rejected measure; Vantage scouting another Trident site.
16 more changed this week. Tracker holds 174 jurisdictions; 114 across 36 states have an adopted pause, ban, rate rule or executive order, and the rest are proposed or pending. The full table ships in the Friday Pro edition.
Your Docket Room Pro seat is included with Dispatch Pro. Track any bill above with one click once you are signed in. Need your code again? Reply to any issue.
The working layer under this briefing is The Docket, in Dispatch Pro
What moved in all 50 states since the last issue, which AI bills actually matter this week, and exactly what they would make you do, by when: obligations, deadlines and penalties pulled from the bill text, similar bills in other states, every hearing and effective date we have reported, plus a Docket Room Pro seat ($49/mo on its own).
You just read today's Docket, open to every reader this once. Pro subscribers get it every morning, with the full Deadline Board and Moratorium Tracker on Fridays.
$29/mo or $290/yr after the trial, less than the Docket Room seat alone. Cancel anytime. Team licenses on request.
Digitize Dispatch is read by lobbyists, government affairs, policy, and regulatory professionals tracking AI legislation, regulation, and policy developments. Put your brand in front of the people shaping how organizations respond to AI policy.
Interested in advertising with us or have any feedback? Send us an email.
