Search the archive

rest. Download GitKraken's guide for engineering managers

AI WON'T KILL THE SOFTWARE BUSINESS, JUST ITS GROWTH STORY

The persistent belief that AI will destroy the software business could be more damaging than the actual impact the technology will have on the industry.

I MISS THINKING HARD

Vibe coding is great for building, but it drastically cuts down the time it takes to come up with creative solutions for technical problems, leading some developers to feel like something is missing.

UPDATE: AI OVERVIEWS REDUCE CLICKS BY 58%

It seems categorically true that AI Overviews siphon away the majority of the clicks once available to top-ranking pages.

IN PRAISE OF THE STACKED PULL REQUEST

Stacked pulled requests are PRs that build on top of each other.

MOBILE CARRIERS CAN GET YOUR GPS LOCATION

Cellular standards have built-in protocols that make your device silently send GNSS locations to carriers.

GIFT CULTURE AND THE INTELLIGENCE THRESHOLD

We create bridges to grow new possibilities and acceptance by giving gifts to each other.

SEE WHY DEVS LOVE GREPTILE

✅ Trusted by engineering teams at NVIDIA, SCALE AI, AND BREX.

✅ Now integrated with Claude Code: install via /plugin.

GOOGLE'S GEMINI APP HAS SURPASSED 750M MONTHLY ACTIVE USERS

Google Gemini has surpassed 750 million monthly active users. The AI chatbot has quickly become a prominent player in the AI space. It had 650 million monthly active users last quarter. In comparison, MetaAI has reported nearly 500 million monthly users, and ChatGPT had an estimated 810 million monthly active users in late 2025.

META SAYS ITS NEXT-GEN AI MODEL, AVOCADO, IS ITS “MOST CAPABLE PRE-TRAINED BASE MODEL TO DATE”

Meta's Avocado model reportedly outperforms top open-source base models. Meta is seeing 10x compute efficiency wins with the model on text-related tasks compared to Maverick and over 100x efficiency gains over Behemoth, a version of Llama 4 that the company delayed last year. The model has now finished pretraining.

INSIDE THE CODEX APP SERVER

OpenAI detailed the architecture of the Codex App Server, a bidirectional JSON-RPC API layer powering Codex across platforms, offering tips for integrating coding agents into real workflows.

KIMI K2.5

Kimi K2.5 is a solid model and exceeds expectations. It is likely now the leading open weights model. The model is excellent given its price. It is an excellent choice for developers who can't afford or don't want to pay for Opus 4.5 and have to go with something cheaper to run their OpenClaw.

AI IS FINALLY EATING SOFTWARE'S TOTAL MARKET: HERE'S WHAT'S NEXT

AI's rise is shrinking the software market, causing unpredictability for SaaS companies. Key players like Salesforce and SAP are integrating AI functionality into popular platforms to adapt, while weaker firms face potential obsolescence. Success hinges on owning intent gateways and effectively leveraging AI for outcomes, positioning few companies to thrive amid industry-wide disruption.

credits/month

MISTRAL INTRODUCES VOXTRAL TRANSCRIBE 2

Mistral's Voxtral Transcribe 2 is a pair of next-gen speech-to-text models, including a real-time version with open weights that delivers sub-200ms latency and low-cost, high-accuracy transcription across 13 languages.

CLAUDE AND CODEX ARE NOW AVAILABLE IN PUBLIC PREVIEW ON GITHUB

Claude by Anthropic and OpenAI Codex are now available for Copilot Pro+ and Enterprise customers on GitHub in public preview. Users can start sessions and assign tasks to these coding agents from the web, mobile app, or VS Code without additional subscriptions. All agent actions, like drafting pull requests and task prioritization, can be managed through GitHub's existing infrastructure.

WINDSURF TAB V2: 25-75% MORE ACCEPTED CODE WITH VARIABLE AGGRESSION

Windsurf's Tab feature was very well received, but it wasn't well maintained after launch. The team has now significantly improved the underlying model and context/data engineering pipeline. This has led to direct Pareto improvements across all of Windsurf's metrics, with an average 54% increase in characters per predict. Tab now uses variable aggression to tailor the experience to each user's preferences.

OWNING A $5M DATA CENTER

comma has been running its own data center for years. All of the company's model training, metrics, and data live inside its own data center in its own office. This post describes how comma's data center works. It is aimed at inspiring other companies to build their own data centers too.

WHY NVIDIA BUILDS OPEN MODELS WITH BRYAN CATANZARO

Nvidia's focus on building open models, spearheaded by efforts like the Nemotron 3 Nano release, positions it uniquely to enhance both its product offerings and the AI ecosystem. Open model development aids Nvidia in understanding compute-heavy AI workloads, which informs their hardware and software design, while fostering greater community collaboration. Nvidia distinguishes itself by emphasizing open data and models to drive AI infrastructure, benefiting businesses without AI acting as an overarching monopolistic event.

C.AI LABS

c.ai labs is a space for rapidly testing creative AI entertainment ideas.

CLAUDE WILL REMAIN AD-FREE

Anthropic announced that Claude will not show ads or include sponsored content in conversations, aiming to preserve trust and integrity in AI interactions, especially in sensitive or complex contexts.

NEMOTRON COLEMBED V2 FOR MULTIMODAL RETRIEVAL

ColEmbed V2 is a family of late-interaction embedding models that achieve state-of-the-art results on ViDoRe benchmarks by improving multimodal document understanding and retrieval.

MANUS EVOLVE

Manus Evolve, powered by Claude Code and Codex CLI, integrates with tools like GitHub and Slack using the EVOLVE SDK.

THE OTHER LEVERAGE IN SOFTWARE & AI

Funds are showing distress - any wobble in expectations will be amplified by borrowing.

BITCOIN SLIPS BELOW $71,000 AS AI-DRIVEN TECH ROUT WORSENS

Bitcoin fell 7.5% in Asian trading on Thursday, slipping below $71,000 to lows near $70,700 as a global technology stock selloff spilled into crypto markets, driven by mounting concerns over peaking AI investment, stretched valuations, and disappointing earnings from firms including Alphabet, Qualcomm, and Arm. Analysts noted that the move below the low-$70,000s has accelerated a broader deleveraging, flushing out crowded positioning built during the post-ETF rally with heavy liquidations and price action now driven more by balance-sheet mechanics than narrative flow, marking the end of complacency rather than the end of institutional participation. The crypto selloff was compounded by brutal commodity moves, including silver plunging 17% and gold falling over 3%, triggering heavy liquidations in tokenized metals products on crypto venues and highlighting fragile conviction across risk assets rather than a clear trend reversal.

TOM LEE: BITMINE'S $6 BILLION ETHER PAPER LOSS IS “BY DESIGN"

BitMine Immersion chairman Tom Lee defended the company's more than $6 billion in unrealized losses on its 4.24 million ether holdings (down from nearly $14 billion in October to approximately $9.6 billion), arguing the drawdown reflects the design of its Ethereum treasury strategy rather than execution failure. Lee positioned BitMine as structured to track and ultimately outperform ether over a full market cycle, similar to an index-style product, comparing paper losses during crypto downturns to market declines in traditional index funds and questioning why similar scrutiny isn't applied to those vehicles. The firm has continued accumulating ETH, adding over 40,000 coins shortly before the latest decline, with its strategy centered on long-term accumulation and staking yield rather than short-term price timing, mirroring approaches used by bitcoin-focused treasury companies that view volatility as the cost of maintaining long-duration core asset exposure.

Y COMBINATOR STARTUPS NOWABLE TO RECEIVE THEIR FUNDING IN STABLECOINS

Y Combinator startups can receive their $500,000 funding in USDC stablecoins instead of cash beginning in Spring 2026. This shift targets faster, cheaper cross-border transactions. Led by partner Nemil Dalal, the initiative leverages improved regulatory clarity from the GENIUS Act to modernize financial infrastructure for founders.

AI NEEDS BLOCKCHAINS — ESPECIALLY NOW

AI systems are breaking the human-scale internet by making it increasingly cheap to generate indistinguishable content and impersonate humans at scale, creating an urgent need for blockchain-based infrastructure to separate humans from machines while preserving privacy and usability. AI needs blockchains for raising the cost of impersonation through decentralized proof-of-personhood systems that make it easy to be one participant but persistently hard to be many (like World ID), creating censorship-resistant decentralized identity systems where users rather than platform gatekeepers control access, enabling portable universal "passports" for AI agents to operate across multiple ecosystems without platform lock-in, and facilitating machine-scale payments through micropayment infrastructure and programmable smart contracts that traditional financial systems cannot handle. Blockchain-based systems paired with zero-knowledge proofs can enforce privacy by default, allowing users to prove specific facts (like age eligibility) without revealing underlying data, solving the paradox where security systems collecting more data make AI impersonation easier.

WHY VISA, MASTERCARD, AND STRIPE AREN'T THREATENED BY CRYPTO

Traditional payment giants are aggressively acquiring crypto infrastructure rather than competing against it: Stripe bought Bridge for $1.1B plus Privy, Visa partnered with Mercuryo for real-time crypto-to-fiat off-ramping through Visa Direct, and Mastercard pursued Zerohash for up to $2B alongside exploratory talks with BVNK for stablecoin payment infrastructure. These acquisitions focus on embedding blockchain rails into existing PSP stacks and compliance frameworks rather than launching standalone token networks, with stablecoins prioritized for instant, predictable value transfer. The next 18 months are expected to bring surging stablecoin volume, regulatory clarification, continued M&A activity, and crypto payments appearing as standard checkout options alongside cards and BNPL.

THE PAYMENT LAYER FOR AUTONOMOUS AGENTS

AI agents require new payment infrastructure to execute autonomous transactions, with incumbent financial institutions competing against crypto-native stablecoin wallets with programmable spending limits. Crypto solutions enable micropayments at $0.001-$0.05 per query with microsecond settlement onchain, positioning them for agent-to-agent commerce and data marketplaces. The AI agents market is projected to grow from $7.6B in 2025 to $180B by 2033, with BCG forecasting 45% annual growth in agentic commerce through 2030.

L2S NEED A NEW ROLE

Vitalik Buterin says the original rollup-as-branded-shard roadmap no longer fits. The Ethereum layer 1 is scaling, and many layer 2s (L2s) won't or can't be true shards, so L2s should pick clear non-scaling value adds, be at least stage-1 when handling ETH, and maximize interoperability. Furthermore, Ethereum should add a native rollup precompile for ZK-EVM proof verification to enable safe, trustless L1 to L2 (and vice versa) interop and synchronous composability.

FIXED RATES ARE THE MISSING PIECE FOR INSTITUTIONAL ON-CHAIN CREDIT

DeFi's money markets currently favor lenders but fail borrowers, especially institutions, that need predictable, fixed borrowing costs, which is why fixed-rate capability is the top priority across Morpho, Kamino, and Euler roadmaps. P2P fixed-rate lending (clean but 1:1 capital-inefficient and liquidity-fragile) and rates markets/swaps (240–500x more capital efficient by hedging floating exposure into fixed), and protocols are building both plus hedging tools to avoid withdrawal/liquidity risks for curators. Together, fixed-rate primitives and rates markets could unlock institutional credit, RWAs, and consumer “borrow-and-spend” products onchain, turning existing lending liquidity into a two-sided, scalable credit market.

TETHER SCALES BACK $20 BILLION FUNDING AMBITIONS AFTER INVESTOR RESISTANCE

Stablecoin issuer Tether lowered its $20 billion funding target to $5 billion following investor pushback on a $500 billion valuation.

$9 BILLION BITCOIN SALE BY GALAXY CLIENT REIGNITES QUANTUM THREAT DEBATE

A Satoshi-era investor sold $9 billion in bitcoin via Galaxy Digital, citing quantum computing risks.

DERIVE OPTIONS EXCHANGE

Derive is an onchain options exchange that has processed $20B in volume whose thesis is that options provide superior leverage to perpetuals by eliminating liquidation risk and funding rate uncertainty.

WHITE HOUSE STABLECOIN YIELD DEADLINE

At a White House meeting between crypto and bank reps, the administration gave an end-of-February deadline to reach a deal on stablecoin yield, with banks pushing to prevent companies like Coinbase from offering rewards on stablecoin holdings.

TINDER LOOKS TO AI TO HELP FIGHT ‘SWIPE FATIGUE' AND DATING APP BURNOUT

Tinder is testing a new AI-powered feature called Chemistry to address swipe fatigue and dating app burnout by learning users' interests through questions and optional access to their Camera Roll. The feature aims to replace endless swiping with a smaller set of more relevant matches as Tinder faces declining sign-ups, lower engagement, and subscriber losses. Chemistry is part of a broader strategy that includes AI-driven recommendations, enhanced safety tools, and increased marketing spend to improve relevance, trust, and appeal among Gen Z users.

LEAKER SAYS IPHONE 18 PRO KEEPS IPHONE 17 DESIGN

Apple is expected to keep the iPhone 17 Pro design for the iPhone 18 Pro, focusing less on a visual refresh and more on performance after strong market reception. The biggest upgrades are likely the A20/A20 Pro 2nm chips for better speed, efficiency, and on-device AI, with possible front-facing tweaks like a smaller cutout or under-display Face ID.

FIGMA ACHIEVES C5 ACCREDITATION

Figma has achieved C5 accreditation from Germany's Federal Office for Information Security, strengthening cloud security and compliance for customers in Germany, Austria, and Switzerland. The accreditation confirms Figma meets rigorous standards for information security, risk management, and operational transparency, making it easier for organizations in the government and finance sectors to use the platform. This milestone builds on Figma's investment in the DACH region, where nearly 90% of DAX 40 companies already use the design collaboration platform.

DESIGN SYSTEMS TEAM STRUCTURE IN 2026

Design systems team structures have changed dramatically due to AI capabilities, with tasks that previously required three people in 2023 now often manageable by a single person with the right AI toolkit. Teams are struggling not with headcount but with leverage, as AI can now handle non-judgment work like documentation, auditing, and scaffolding while humans focus on strategic decisions and relationship building. The key is implementing systems that create scale before adding headcount, as teams that understand what to automate versus what to own can outperform much larger teams stuck in inefficient workflows.

A DESIGN CONVERSATION AT THE APPLE CREATOR EXPERIENCE

A designer shared his experience at Apple's LA Creator Experience, discussing how tools like Pixelmator Pro and iPad help overcome the fear of starting creative projects. After losing all his work in a 2006 robbery, he started the blog Abduzeedo to document his process and has been using Pixelmator since 2007. He emphasized that good design tools matter because they help creators push through initial creative barriers and start new projects.

AI RUNS ON TEXT. SO SHOULD YOU

AI tools like Claude Code are powerful not because they're flashier, but because they work directly with your existing files, making plain text the shared language where human thinking and AI collaboration really scale. The quiet advantage is externalizing your thinking in accessible text, letting AI synthesize, reorganize, and build on it over time without complex systems.

COMPARE AND PICK FONTS

Font Mojo is a tool for comparing and selecting fonts from different sources. It allows users to browse fonts with sorting and preview options.

PROFESSIONAL COLOR SYSTEM BUILDER

Create accessible palettes with perceptual accuracy, export to 10+ formats, including Tailwind & Figma.

ANIMATE SVGS

Stroke lets you create hand-drawn SVG animations for logos, signatures, and illustrations.