Your signal. Your price.
Dylan Field says Figma, founded in August 2012, helps teams design and build software. Design is its core, but people also use Figma for marketing, advertising, presentations, whiteboarding, and other visual work.
Field argues that as AI makes coding easier, design, brand, marketing, and a clear point of view become stronger sources of product differentiation. He calls design the new code, while stressing that Figma still needs engineers.
Field rejects both blanket claims that SaaS is safe and that SaaS is dead. He points to Cursor's growth as evidence that a company can compete with frontier labs by obsessing over users, anticipating their needs, and building the right product before it has its own model.
A VPS provider’s server hosting Max’s website died while its status page remained green; support offered a downtime refund but no server recovery. Max restored the site from mostly current backups and said AI helped rebuild the missing pieces.
Google Play Store updates will allow Android users to vibe code custom applications directly on devices and share them within trusted friend networks. Adam Curry demonstrated building a functional, custom podcast discovery application using a single text prompt.
Emmanuel Ameisen argues that scaling compute and text richness yields predictable empirical improvements under scaling laws. Product breakthroughs occur when smooth capability gains cross discrete reliability thresholds, such as increasing code generation accuracy from 80 percent to 99 percent.
EliseAI focuses technical recruitment in San Francisco and New York, seeking software engineers and academic researchers. Mario Martone targets candidates with strong theoretical backgrounds who want to apply scientific rigor to production software.
David Sacks said coding assistants are expanding software creation for bespoke enterprise needs, while Ben Horowitz expects software prices to fall as total software use grows. Both argued that coders remain necessary to oversee and maintain AI-generated systems; Horowitz pointed to data, user feedback, and product quality as durable advantages.
Jones is building an ad-removal specification that must preserve episode content exactly; he will not publish it until it works reliably across a test set. His use case includes blocking repetitive or irrelevant ads without removing every advertisement from a show.
Cognition is testing dynamic Devin token limits: its system estimates equivalent human work, grades engineers’ usage, and raises limits for efficient sessions. Devin Coach also flags common mistakes, including vague requests and using costly models for routine tasks; developers criticized the approach as factory-style management.
YouTuber Max Blade and livestream viewers built Kong Royale by directing AI agents to select chat-submitted changes overnight for morning builds. Blade says Opus 5.5 and Claude Max accounts made the project workable in a few weeks, extending the earlier idea of vibe-coding complete games.
Meta’s Muse Spark Open has 816 billion total parameters, 42 billion active parameters, and a million-token context window. Meta claims open-state-of-the-art results on Jobbench and MCP Atlas; Artificial Analysis scored it 39, behind Kimi K3 on coding tests but ahead of Nemotron 3 Ultra and Inkling.
Blitzy says one customer built a new application with 534,000 lines of code, compressing a 65-week roadmap into two weeks; another shipped an application without a front-end engineer. The product autonomously plans, builds, tests, and validates entire greenfield applications.
A host said Indian students are protesting because AI threatens the IT and software careers many viewed as a route out of poverty; the protests have been associated with a satirical group called the Cockroach Party.
SF built the current jukebox with AI after finding the first bot difficult to maintain and local coding agents unreliable. SF used donation funds for an Anthropic subscription, but said the AI-generated system remains a black box and its long-term value is uncertain.
Mert joined the Near Foundation Council, which David reads as a technical endorsement of Near’s privacy work and a potential fit with Zcash; the hosts note that Starknet already uses its own Cairo language and is not an EVM chain.
Naous says companies were spending 5% to 20% of infrastructure budgets on observability as a proxy for reliability. Microservices increased telemetry volume and cost, while each extra observability dollar produced less reliability than before.
Naous says AI widens the gap between strong and average engineers: experienced engineers can break down complex work and direct agents, while weaker engineers do not automatically become strong. He sees model and coding-harness differences, but also transferable ways to work with AI.
Naous uses Claude Code from a Windows machine through WSL, with Linux command-line workflows. He coordinates three or four sessions, has agents review design documents and pull requests, checks algorithms and tests himself, and then assigns parallel implementation work.
Naous says AI helped him ship a complex data-lake auto-promotion feature in about 10 days, compared with an estimated six months for one engineer before. The feature turns fields in semi-structured log data into independently queryable columns.
Naous argues that AI can write production code without independently understanding distributed-system behavior. Engineers still need to understand query semantics, correctness, consistency, and workload tradeoffs, then explain the requirements the model cannot infer from code alone.
Naous says AI performs better on operational work when engineers provide benchmarks, representative queries, and explicit constraints. It can identify query optimizations once performance is measurable, but it cannot reliably choose system memory settings without experiments on the running system.
Grepr’s troubleshooting approach looks for new patterns in telemetry and surfaces early signals such as exhausted connections or memory errors, rather than starting with an alert and backtracking through the system. AI then inspects code, checks whether the signal has appeared before, and tests possible explanations.
Casado argues that code and other digital tasks are easier because their solution distributions resemble their training data; real-world systems bring rare states and unpredictable conditions. He cites self-driving’s long path from the DARPA Grand Challenge to deployments as evidence of the cost and time involved.
Kedrosky argues that token volume can mislead investors because revenue depends on both usage and price, and token prices are falling sharply. He says early demand is skewed toward coders who use expensive models heavily, unlike many consumers and ordinary business users.
Adam Curry says he uses at least eight coding agents and runs five local machines that do not send data to the cloud. He sees personal data files as useful agent context, while acknowledging that a virtual machine rented from Meta raises an ownership question.
Adam Curry says coding agents let people build software around their own needs instead of accepting generic apps. He built Cleanfeed-like functionality for himself and may cancel a $36 subscription; a homeschooling user reported saving three hours weekly and getting a custom design tool built.
Developer Brandon Thomas used Claude to decompile Photoshop and rebuild it through clean-room reconstruction, then generated incomplete copies of Adobe’s suite and offered them free against a $100-per-month product. Nathaniel Whittemore says Adobe shares opened down 3% after the story spread, while arguing that enterprise SaaS customers often pay for maintenance and support, not software code alone.
AI-assisted decompilation and game mashups are colliding with a struggling console business: console sales are at a 13-year low, studios have cut tens of thousands of jobs, and hackers are using AI to speed PS5 exploits. Bloomberg Intelligence argues streaming and data-center AI could weaken consoles; the PS5 base model now costs $650, up from $500 at launch, while the Steam Machine is priced at $1,000.
Game-engine developer George Kahl says Opus 5.5 can handle code, shaders, systems, level logic, sound, tools, and debugging well enough to move toward solo-made premium games; he estimates that individual developers could release their own visions in one to two years. Andrew Curran says models already produce stable game mashups in a day, while Alex Finn predicts a 2027 AI-game gold rush.