Anthropic's Record $81B IPO + Gemini Neon, Figma Agent, Bing AI Ads, Snap AR CLI
This daily tech brief covers Anthropic's record $81.34B Nasdaq IPO, Google's Gemini 4 Neon model with 1M token context and $3/$15 per million token pricing, Figma's Agent in Dev Mode for design-to-code queries, Microsoft's integration of Performance Max ads into Bing AI summaries, and Snap's Easy Lens CLI for building AR lenses via AI coding tools.
Anthropic's Record $81.34B Nasdaq IPO
Anthropic debuted on Nasdaq under ticker ANT at $83 per share. The offering comprised 980 million shares (956 million new, 23.65 million from existing shareholders), valuing the total public offering at $81.34 billion before underwriting fees — the largest IPO ever according to the Associated Press. First-day trading opened at $111.50 and closed at $97.17, a 17.1% gain over the offer price; the article notes that "first-day pop" depends on which reference price is used.
The prospectus revealed revenue of approximately $33.2 billion for the nine months ended September 2026, alongside a net loss that includes large fair-value changes on convertible notes. The author emphasizes that revenue, accounting loss, and operating cash flow each answer different questions for AI companies.
Google Launches Gemini 4 Neon: A Middle Tier for Reasoning and Speed
Positioned between the Hydrogen and Carbon models, Gemini 4 Neon targets coding, knowledge work, and scientific tasks. It supports a 1-million-token context window. Pricing via the standard API: $3 per million input tokens, $0.15 per million cached input tokens, and $15 per million output tokens. Actual costs depend on cache hits and output volume.
Google cites an Artificial Analysis benchmark where Neon scored 61 at its highest reasoning tier, placing third at launch. The score reflects a specific tier and test suite. Access is available to Gemini Pro and Ultra subscribers in the Gemini app, and to developers via the Gemini API; Workspace account rollout follows its own schedule.
Figma Agent Enters Dev Mode: Query Designs Directly
Figma introduced Agent inside Dev Mode. Developers can ask why a layout is set a certain way, compare two versions, view a diff list, or request CSS/React implementation suggestions and asset links — all without switching between chat and design files. Permissions follow existing seat rules: Dev seats can query, while edit permissions remain separate. The author cautions that style and code suggestions still require developer judgment regarding project components, state logic, and accessibility.
Figma also expanded in-file image generation: a single prompt can produce multiple variants from the same model or one variant each from different models for side-by-side comparison.
Bing AI Search Summaries Add Ad Slots
Microsoft announced that existing Performance Max and AI Max campaigns can now appear in Bing's AI-generated search summaries (outside China). Advertisers reuse current campaigns. The shift changes the user reading path: traditional search scans a link list, while AI summaries keep part of the explanation on the results page, and paid content now appears in that flow. Microsoft references purchase-behavior data from prior Copilot interactions; click-through and conversion rates for the new ad slots will only be known after launch.
Snap Releases Easy Lens CLI for AR Development via AI Coding Tools
Snap's free Easy Lens CLI integrates with Codex, Claude Code, and Cursor, letting developers build Snapchat Lenses through natural-language descriptions. It provides over 100 AR components and 50+ AI asset-generation tools. A mask that reacts to head movement — generating assets, attaching tracking, and tuning triggers — can be assembled in one workflow.
The toolchain connects preview, on-device testing, submission, publishing, and performance checks. The author argues that embedding phone validation into the same loop better matches real AR production than merely shortening generation time, because stability of motion tracking, occlusion handling, and cross-device smoothness are critical.
Signed-in readers can open the original source through BestHub's protected redirect.
This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactand we will review it promptly.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.
