Horizonself-hosted digest
← 2026-09-19 2026-09-20

Horizon Daily - 2026-09-20

From 21 items, 3 important content pieces were selected


Technology News
1. California Governor Signs Order Seeking Reporting of AI Loss-of-Control Incidents ⭐️ 7.0/10

Technology Blog
1. Grit Your Teeth and Ship Imperfect Work ⭐️ 5.0/10

Financial News
1. Beijing Regulator Opens Probe Into Meituan, Fliggy, Tongcheng and Tujia Over Hotel Booking Practices ⭐️ 7.0/10


Technology News

California Governor Signs Order Seeking Reporting of AI Loss-of-Control Incidents ⭐️ 7.0/10

California Governor Gavin Newsom signed an executive order on September 19 aimed at strengthening AI safety, including a proposed requirement that companies report AI agent "loss-of-control incidents" and a possible requirement that advanced models carry emergency shutdown mechanisms. The order convenes an expert panel to propose guidance within two months for improving AI safety law, and it proposes periodic audits of AI labs; Newsom said insufficient federal regulation makes state action necessary. These are directives and proposals rather than enacted law, and the account comes from a Telegram post reposted by Sina Finance without primary-source detail.

telegram · zaihuapd · Sep 19, 05:44

「Background」 California's action is an executive order, which directs agencies and expert panels to produce guidance rather than imposing binding requirements itself. POLITICO reported that Newsom issued the order partly to speed up implementation of two AI laws he had recently signed and to explore proposals such as requiring AI companies to build an emergency shutdown mechanism, or "kill switch," for their most advanced models — meaning the two-month guidance deadline and proposed lab audits described in the order are recommendations at this stage, not enacted rules.

「Impact」 AI companies operating in California may face earlier independent oversight and safety checks, because the order directs state action to accelerate implementation of California’s new AI oversight law and to advance an AI “kill switch” framework rather than waiting for new legislation. Developers should watch the expert guidance due within two months and prepare for possible audit and incident-reporting expectations, though the order itself does not yet enact mandatory reporting or shutdown requirements.

References

Tags: #AI regulation, #AI safety, #California, #AI governance, #policy


Technology Blog

Grit Your Teeth and Ship Imperfect Work ⭐️ 5.0/10

rss · Sean Goedecke · Sep 20, 00:00

「Background」 In this personal essay, Sean Goedecke argues that being good at building and being good at shipping are separate, even countervailing skills: strong taste in creative work makes people reluctant to publish work that disappoints them.

「Solution」 He cites Ira Glass's taste-gap idea and applies it to programming, where gifted engineers' desire for elegant, correct systems makes them freeze up, retreat into safe small domains, or spin in shame rather than ship. In large codebases, flaws are inevitable and consistency is the priority, so the author argues it is sometimes right to duplicate non-catastrophic flaws; a bad diff can be improved with time, but no diff cannot. On writing, he says he often finishes a draft without thinking it is very good, yet mostly publishes anyway. Shipping improves with practice: after a month without posting, the next draft feels too weak to publish, while daily posting makes each draft feel fine. Rereading old posts, he cannot tell which he felt good about, and sees no correlation between that feeling and popularity; he lists posts he disliked that resonated and posts he liked that did not. Because he cannot predict what readers find useful, high-volume output yields better than a few highly polished pieces, and he advises not being precious about ideas, noting he has written roughly thirty posts about shipping.

「Takeaway」 The author concludes that since you cannot control which piece succeeds, you must be momentum-based rather than outcome-based: grit your teeth and ship.

Tags: #shipping, #perfectionism, #software engineering culture, #writing, #developer productivity


Financial News

Beijing Regulator Opens Probe Into Meituan, Fliggy, Tongcheng and Tujia Over Hotel Booking Practices ⭐️ 7.0/10

Beijing's market regulator said on September 19 that it has opened an investigation into online travel platforms Meituan, Fliggy, Tongcheng and Tujia over alleged illegal practices in hotel accommodation, including paid bidding for traffic display and requiring merchants to sell at the lowest price across the whole internet, which would strip them of pricing autonomy. The four companies said they are cooperating with regulators; no findings or penalties have been announced yet.

telegram · zaihuapd · Sep 19, 07:47

「Background」 China's online travel sector already has a precedent: in July 2026 the State Administration for Market Regulation fined Ctrip a total of 5.179 billion yuan (1.658 billion yuan in confiscated illegal gains plus 3.521 billion yuan in fines) for abusing market dominance, the industry's first antitrust case. According to Tencent News, Beijing's market regulator had already summoned and issued administrative guidance to Ctrip, Qunar, Meituan, Fliggy and other platforms in March 2026, together with the city's commerce and culture-and-tourism bureaus.

「Impact」 Hotels and other accommodation merchants listing on these four platforms are the businesses most directly affected, because the practices under investigation — paid bidding for traffic placement and requirements to sell at the lowest price available anywhere online — shape how their rooms are priced and how visible they are to customers.

References

Tags: #China regulation, #online travel platforms, #algorithmic marketing, #hotel accommodation, #pricing restrictions