The federal government brought a list of roughly 20,000 names to a California courtroom and asked for an extraordinary election-day consequence: separate any mail ballots returned by those voters and hold them out of the count while their citizenship was checked. On Wednesday, October 28, U.S. District Judge David Carter denied the requested temporary restraining order. California election officials may continue their existing ballot procedures as the November 3 midterm election approaches. The ruling decides the emergency request, not the underlying citizenship of every person on the list or the final lawsuit. [California's account of the ruling](https://oag.ca.gov/node/632032) and [the Associated Press report](https://apnews.com/article/481bed64919ad193e80e051d28beb2c0) describe the decision.

That distinction matters because the administration's claim and its requested remedy were different propositions. Federal officials said the listed people were likely noncitizens. They sought a court order that would change how election workers handle ballots already arriving. The judge found the requested intervention would disrupt an election in progress and risk excluding eligible voters. A court can take a claim seriously without granting a last-minute order that treats the claim as proved.

The record also raises a practical problem. Once election workers separate an accepted ballot from its identifying envelope, the secret ballot cannot simply be traced back to a voter. The government cited that feature as a reason to act quickly. California argued that changing procedures midstream would itself create confusion and threaten lawful votes. The judge accepted the urgency of that second risk. The state says federal officials acknowledged that their list could include citizens; its attorney general also says officials have not supplied enough underlying data to test the entire list. Those are the state's assertions about the evidence, not a finding that every name is wrong. [California's October 28 statement](https://oag.ca.gov/node/632032) sets out its position.

A preliminary Los Angeles County review of 200 names, described in California's earlier court account, found many inactive or canceled registrations and some records listing a U.S. birthplace. That sample is useful as a warning about list quality. It is not a full audit of 20,000 people, and a birthplace field alone cannot settle every citizenship question. Nor does a registration entry establish that a ballot was cast. The [state's October 26 account](https://oag.ca.gov/news/press-releases/attorney-general-bonta-scores-early-legal-victory-against-trump-administration%E2%80%99s) gives the sample details.

The legal path has already been unusual. Earlier this week, the Ninth Circuit declined the government's attempt to obtain emergency relief before the district court had ruled. Wednesday's decision now gives that district court answer. The case may continue, and any appeal would be a new event. Neither the appellate refusal nor the temporary-order denial is a certified account of how many ineligible voters, if any, are on California's rolls. [The Ninth Circuit's opinion](https://cdn.ca9.uscourts.gov/datastore/opinions/2026/10/26/26-7070.pdf) explains the earlier procedural step.

There is a larger question here than the fate of one list. Election administration depends on checking eligibility before a ballot is counted and preserving the secrecy of a lawful vote afterward. Those duties collide when a government brings an unverified list days before an election and asks a court to stop processing thousands of ballots. The judge chose continuity while the evidence is tested. That is a consequential decision, but it is a narrow one. The public still deserves a clear account of how the names were assembled, how many match active registrations, and what investigation follows. Until then, 20,000 is the size of an allegation, not a count of illegal votes.

Israel's October 27 election has produced a strong signal and no settled government. As counting continued Thursday morning, October 29, neither Prime Minister Benjamin Netanyahu's allies nor the parties trying to replace him appeared to command a majority in the 120-seat Knesset. The final result is still pending. The next contest is political: which parties will agree to govern together, and on what terms? [The Associated Press](https://apnews.com/article/9a0620a1314fdc127245ba1a7f0aa762) reported the unresolved balance early Thursday.

Arab-led parties have become central to that question after a strong showing in the partial count. Opposition figures could have a route to power if they can secure their support, but a mathematical route is not a coalition agreement. The parties differ sharply on security, representation and the terms of participation. Some potential partners have resisted depending on Arab-led parties. A voter can see the contradiction: parties that need seats may still refuse to treat the people who won them as legitimate partners. [AP's account of the opposition's options](https://apnews.com/article/da107a1b69b5cf763394bae89ad759bc) describes that problem.

The count also needs careful language. Interim seat estimates can change as remaining ballots are included. Election-night projections, partial official figures and final allocations are different things. The Central Elections Committee says final results are due by November 4. The president's consultations with parties and the formation of a government follow the official result, not a television graphic. [The election committee's results site](https://cdn-votes26.bechirot.gov.il/) states the final-publication date.

A fresh election would carry its own uncertainty. The voters who turned out this week, including citizens who traveled to vote, may not assemble in exactly the same numbers again. Campaigns therefore have an incentive to present coalition talks as impossible or inevitable before testing them. Readers should watch for a named agreement, a seat count based on final returns and a formal mandate to try to form a government. Until then, the headline is an opening negotiation.

The Federal Reserve held its target interest-rate range at 3.75% to 4% on Wednesday, October 28. The decision passed 10 to 2. The unusual point is the direction of dissent: Lisa Cook and Beth Hammack preferred a quarter-point increase. The committee's own statement describes economic activity as strong and inflation as elevated. [The Federal Reserve's statement](https://www.federalreserve.gov/newsevents/pressreleases/monetary20261028a.htm) gives the vote and the policy range.

A hold can sound like reassurance, especially six days before a national election. It is more precise to call it a decision to wait. The committee did not say inflation had returned to its 2% goal. It said it needs more information to judge the appropriate stance. The two dissenters judged the current evidence strong enough to tighten now. Their disagreement is over timing and risk, not over whether inflation matters.

The target range is the rate-setting committee's policy instrument. It is not the rate a family receives on a mortgage or a business pays on a loan. Those prices reflect several forces, including expectations about future policy and broader credit conditions. The Fed's implementation note specifies a 3.90% rate on reserve balances effective October 29, a separate operational setting that helps keep market rates within the target range. [The implementation note](https://www.federalreserve.gov/newsevents/pressreleases/monetary20261028a1.htm) spells out the mechanics.

The next scheduled policy meeting is December 8–9, according to the Fed's calendar. By then, policymakers may have more evidence about spending, jobs and prices. Today's vote therefore leaves a live question for December: whether the majority's patience survives another round of inflation data. The dissenters have made clear which side of that question they occupy. [The Fed calendar](https://www.federalreserve.gov/monetarypolicy.htm) lists the next meeting.

Microsoft reported $92.3 billion in revenue for the quarter ended September 30, 2026, up 19% from a year earlier. Azure revenue rose 48%, and Microsoft Cloud revenue reached $63.9 billion, up 30%. Those are substantial operating gains. The company also reported $36.6 billion in GAAP net income, up 32%. The numbers belong to its first quarter of fiscal 2027, released October 28. [Microsoft's earnings release](https://www.microsoft.com/en-us/investor/earnings/fy-2027-q1/press-release-webcast) contains the period and figures.

The profit comparison needs a second sentence. Microsoft says investments in OpenAI added $207 million to this quarter's GAAP net income. In the comparable quarter a year earlier, those investments reduced GAAP net income by about $3.1 billion. Excluding that effect under the company's definition, adjusted net income rose 18%, to $36.4 billion. The 32% GAAP increase is real as reported; it also includes a favorable swing from a prior-period investment loss. Calling the entire increase an acceleration in the core business would overread it.

The release offers another number likely to attract attention: commercial remaining performance obligation reached $677 billion, up 73%. That is a measure of contracted business still to be recognized as revenue over time. It is not current-quarter sales and does not tell readers, by itself, when all the services will be delivered or what margin they will earn. The better question is how rapidly those commitments turn into revenue and cash while the company keeps funding the infrastructure to serve them.

There are counterweights in the same release. Devices and Consumer revenue fell 3%, while Windows OEM and Devices revenue fell 22%. Microsoft also returned $10.2 billion to shareholders through dividends and repurchases. The quarter is strong, but strength is not a license to merge unlike measures. Revenue, investment gains, contractual backlog and cash returns each answer a different question. The company has supplied the pieces; investors should keep their labels attached.

WM's third-quarter 2026 revenue rose 6.3% to $6.846 billion, according to results released October 27 for the three months ended September 30. Collection and Disposal volume, however, declined 0.4%, or 0.5% after the company's workday adjustment. WM said core price increased 5.8%. It also cited higher energy surcharges and more volume in recycling and renewable-energy businesses. [WM's earnings release](https://investors.wm.com/news-releases/news-release-details/wm-announces-third-quarter-2026-earnings) provides the measures and their definitions.

That is a useful antidote to the lazy equation of higher sales with more work. A company can increase revenue by collecting more material, charging more for the same service, changing its mix or adding surcharges. WM's disclosure says pricing played a large role this quarter. Management also says some lost collection volume reflects its decision to shed lower-margin residential business. That may improve returns, but investors should still distinguish a deliberate exit from demand growth.

The margin story is similarly mixed. WM reported adjusted operating EBITDA of $2.121 billion, compared with $2.005 billion a year earlier. Its adjusted operating EBITDA margin slipped from 31.1% to 31.0%. The company says favorable pricing and improved healthcare operations helped, while higher fuel and risk-management costs offset those gains. The adjusted measure is defined by management and differs from GAAP income from operations; both appear in the release.

Cash conversion looks stronger on the longer period. WM said free cash flow for the first nine months of 2026 was $3.01 billion, up from $2.11 billion in the comparable 2025 period. It returned more than $2.72 billion to shareholders in those nine months. WM's free-cash-flow definition is a non-GAAP measure, so it deserves the same careful label as its margin. The interesting business story is clear without a heroic claim: the company can raise prices in a difficult cost environment, but it still has to show how durable that power is when volume is soft.

Two films that reached French cinemas on October 28 approach pressure from opposite directions. László Nemes's Moulin follows the final ordeal of Resistance leader Jean Moulin after his arrest in 1943. Marion Le Corroller's debut Sanguine, also titled Species, follows an emergency-room intern whose body begins to change under stress. One draws on occupation and interrogation; the other turns workplace exhaustion into body horror. Cannes lists both releases in its current French-cinema program. [The festival's release listings](https://www.festival-cannes.com/en/) and [film pages for Moulin](https://www.festival-cannes.com/en/f/moulin/) and [Sanguine](https://www.festival-cannes.com/en/f/sanguine/) supply the basic details.

The directors' descriptions point to a shared formal concern. Nemes told the Cannes festival he wanted viewers close to Moulin's experience rather than at the safe distance of a conventional biography. He described a film governed by surveillance and uncertainty. Le Corroller described Sanguine as a satire of burnout, using the disquiet of body horror to make an overburdened worker's condition visible. These are the filmmakers' accounts of their aims, not a verdict on what either film achieves. [Cannes's Nemes interview](https://www.festival-cannes.com/en/2026/laszlo-nemes-i-wanted-the-viewer-to-be-one-of-jean-moulins-comrades/) and [its Le Corroller interview](https://www.festival-cannes.com/2026/rencontre-avec-marion-le-corroller-nouvelle-figure-du-body-horror-a-la-francaise/) explain the choices.

That pairing makes an inviting cinema question. A historical drama can turn a famous name back into a vulnerable person. Horror can take an ordinary institutional demand and render it physically grotesque. Both methods ask the audience to feel pressure before translating it into a thesis. They also carry different hazards: intimacy can narrow historical context, while metaphor can flatten the real work of care into spectacle.

No review can be written from a synopsis and a director interview. What the release calendar does offer is a reason to watch closely. The striking feature is not merely that two Cannes titles opened on one day. It is that both filmmakers treat the body as the place where an institution's power stops being abstract.

The California ballot case should leave one simple civic rule behind: the power to challenge a vote comes with a duty to explain the challenge. On October 28, a federal judge denied the administration's request to set aside ballots associated with roughly 20,000 people it described as likely noncitizens. The court did not settle every eligibility question. It refused the emergency remedy sought on the evidence and timing presented. [The state's account of the order](https://oag.ca.gov/node/632032) and [AP's report](https://apnews.com/article/481bed64919ad193e80e051d28beb2c0) establish the immediate result.

The government has a legitimate interest in preventing unlawful voting. Citizens have an equally serious interest in casting a ballot without being singled out by a poorly explained database match. These interests do not have to be performed as opposing slogans. They can be tested against a method. Who assembled the list? What records were matched? How were false matches checked? How many names represent active registrations? How many people actually returned ballots? A transparent answer would help election officials investigate real problems and let the public judge the scale of the claim.

Emergency litigation is a bad place to discover that the method cannot be explained in time. The state's preliminary review of 200 names raised reasons to question the list's accuracy, though it did not audit all 20,000. That should prompt more scrutiny, not a victory lap claiming every listed voter is eligible. [California's earlier account](https://oag.ca.gov/news/press-releases/attorney-general-bonta-scores-early-legal-victory-against-trump-administration%E2%80%99s) describes the sample.

My view is that election authorities should publish enough about any such process for independent examination while protecting personal voter data. Investigate specific evidence, correct genuine errors and prosecute proved violations. The public deserves both an accurate roll and a ballot system that does not convert suspicion into mass exclusion. When the requested remedy is unusually broad, the explanation must be unusually good.

Microsoft's latest earnings release contains a number built for a giant headline: $677 billion in commercial remaining performance obligation. It is up 73% from a year earlier. The figure describes contracted business that has yet to be recognized as revenue. It says something important about demand. It does not say that $677 billion arrived in the September 2026 quarter, or that its future profit is already known. [Microsoft's October 28 release](https://www.microsoft.com/en-us/investor/earnings/fy-2027-q1/press-release-webcast) provides the number.

The temptation is understandable. In an AI boom, a contracted pipeline sounds like proof that every data center will pay for itself. Yet a contract, a delivered service and retained cash occupy different points in time. Hardware and power must be obtained before some future revenue can be earned. Customers may also change the pace at which they use a service. None of this means the backlog is weak. It means the backlog is the start of an accounting story, not its end.

Microsoft has earned attention for operating growth. Azure revenue rose 48% year over year in the quarter, and total revenue rose 19%. Those figures are a much firmer description of what has already happened. Even the profit headline needs care: a prior-year OpenAI investment loss makes the GAAP growth rate look much stronger than the company's adjusted comparison. The point is not to dismiss growth. It is to avoid combining a present sale, an investment mark and a future commitment into one triumphal number.

I would like companies selling the AI future to make a habit of explaining how contracts convert into revenue, capacity and cash over successive quarters. Investors can then reward execution instead of scale alone. A huge backlog may eventually justify a huge buildout. Today's release cannot settle that question. It gives readers the first column of a ledger they should keep open.