Search the archive

POSTGRESQL 18 RELEASED!

PostgreSQL has a new I/O subsystem that improves performance for workloads of all sizes. Other features include virtual generated columns, a database-friendly uuidv7() function, and easier integration with single-sign-on systems. The release makes major-version upgrades less disruptive. Details about the new features are available in the article.

INTRODUCING PERPLEXITY SEARCH API

The Perplexity Search API provides access to the same global-scale infrastructure that powers Perplexity's public answer engine. It allows developers to tap into information across the internet with one simple yet powerful interface. Unlike other API offers, it provides rich structured responses that are ready for use in AI and traditional applications alike. Perplexity has released an SDK, an open-source evaluation framework, and a deep dive into how its research team designed, optimized, and evaluated the API.

BECOMING A RESEARCH ENGINEER AT A BIG LLM LAB - 18 MONTHS OF STRATEGIC JOB HUNTING

Max Mynter signed as a research engineer with Mistral earlier this week. The road to getting there was not an accident - Mynter had strategically worked towards this outcome for an extended period. This account of his personal experiences shares a few things that worked out for him. Mynter's main tactic was to try to get an information advantage and then act on this information to be prepared when it matters. The preparation can involve strategic high-effort projects or tactical quick-win initiatives.

THERE ARE MORE ROBOTS WORKING IN CHINA THAN THE REST OF THE WORLD COMBINED

China has five times as many robots working in its factories as the US. Until last year, China installed more imported robots in its factories than domestically made ones - last year, nearly three-fifths of the robots installed in the country were made locally. While Chinese companies still lag behind foreign competitors in their ability to manufacture some key components, China has a large number of skilled electricians and specialized computer programmers who can install robots, as well as an artificial intelligence industry that is strongly focused on tracking and improving every aspect of factory equipment performance.

$200 gift

ANTHROPIC CPO ADMITS THEY RARELY HIRE FRESH GRADS AS AI TAKES OVER ENTRY-LEVEL TASKS

Anthropic won't hire fresh graduates because they lack experience, but entry-level jobs that used to give them that experience are increasingly being automated.

GOOGLE DEEPMIND UNVEILS ITS FIRST “THINKING” ROBOTICS AI

Google DeepMind's Gemini Robotic project has announced a pair of new models that work together to create the first robots that 'think' before acting.

META INTRODUCES VIBES FEED FOR AI-GENERATED CONTENT

Meta's Vibes feature gives users the ability to create and share short-form AI-generated videos.

on Marqeta. Find out why

ANNOUNCING CLOUDFLARE EMAIL SERVICE'S PRIVATE BETA

Cloudflare's Email Sending capability allows users to send transactional emails directly from Cloudflare Workers.

SNAPPY

Snappy is a compression/decompression library widely used inside Google.

OPENAI SAYS GPT-5 STACKS UP TO HUMANS IN A WIDE RANGE OF JOBS

OpenAI's GDPval covers a very limited number of tasks people do in their real jobs, but the company claims that testing on the evaluation shows that GPT-5 is approaching the quality of work produced by industry experts.

🌀 Spin up a Together Instant Cluster

CHATGPT PULSE PREVIEW

ChatGPT Pulse is a new feature for Pro users on mobile that proactively delivers personalized research and updates via visual cards based on user chats, preferences, and connected apps.

PERPLEXITY SEARCH API

Perplexity has released a new Search API that gives developers access to the same large-scale infrastructure behind its public answer engine, enabling high-quality retrieval over hundreds of billions of web pages.

MUSK'S XAI SUES OPENAI, ALLEGING THEFT OF TRADE SECRETS

xAI alleges OpenAI conducted a "coordinated, unfair, and unlawful campaign" to recruit key employees and induce them to bring trade secrets, including xAI's rapid data center deployment methods.

WHAT ARE POPULAR AI CODING BENCHMARKS ACTUALLY MEASURING?

Popular coding benchmarks measure something narrower than what their names suggest. Claude scoring 80% on SWE-bench doesn't translate to it one-shotting 80% of the tasks thrown at it. This post looks at SWE-bench Verified, SWE-bench Pro, Aider Polyglot, LiveCodeBench, and other benchmarks to see what they are actually measuring. Designing a good benchmark is highly labor-intensive. Without human review and annotations, it's nearly impossible to make a good benchmark.

IS OPENAI'S REINFORCEMENT FINE-TUNING (RFT) WORTH IT?

OpenAI's reinforcement fine-tuning (RFT) for o4-mini is supposed to be able to train models to get better at specific tasks using reinforcement learning. It costs up to 700 times more than supervised fine-tuning but only seems to deliver clear wins on agentic coding tasks. The technology gives engineers flexibility in reward design through flexible grader configurations, and the performance gains can be substantial when it works, but it is expensive with limited applicability, making its use difficult to justify.

TOWARD COMPUTATIONAL TASTE: LLMS, AESTHETICS & JUDGMENT

LLMs are being used to model and optimize taste in diverse fields through personalized systems like Taste Engines, Aesthetic LLMs, and Taste Tribes. Methods like LoRe and models like TAPO enable LLMs to adapt to individual aesthetic preferences, potentially revolutionizing recommendation engines and social networks. This shift towards computational taste signifies a new era where machines not only reflect but also shape human preferences, offering both cultural and economic opportunities.

you can stay secure. Read the eBook by Dell and Intel

CONTINUING TO BRING YOU OUR LATEST MODELS, WITH AN IMPROVED GEMINI 2.5 FLASH AND FLASH-LITE RELEASE

Google DeepMind has released updated versions of Gemini 2.5 Flash and 2.5 Flash-Lite. The models, available on Google AI Studio and Vertex AI, feature improvements in quality and speed compared to the current stable models. They are also more efficient, with a 50% reduction in output tokens for Gemini 2.5 Flash-Lite and a 24% reduction for Gemini 2.5 Flash. The models are not intended to graduate to a new, stable version, but will help shape future stable releases.

GEMINI ROBOTICS 1.5 BRINGS AI AGENTS INTO THE PHYSICAL WORLD

Google DeepMind has released two models that unlock agentic experiences with advanced thinking. Gemini Robotics 1.5 is a vision-language-action model that turns visual information and instructions into motor commands for a robot to perform a task. The model thinks before taking actions and shows its process and also learns across embodiments. Gemini Robotics-ER 1.5 is a vision-language model that reasons about the physical world, natively calls digital tools, and creates detailed multi-step plans to complete a mission. These advances will help developers build more capable and versatile robots that can actively understand their environments and complete complex multi-step tasks in a general way.

GDPVAL: REAL-WORLD AI BENCHMARKING

OpenAI's GDPval is an evaluation benchmark that tests model performance on real-world, economically valuable tasks across 44 occupations.

BECOMING A RESEARCH ENGINEER AT A BIG LLM LAB - 18 MONTHS OF STRATEGIC JOB HUNTING

Max Mynter signed on as a research engineer with Mistral earlier this week after years of working towards the outcome strategically. This account of Mynter's personal experiences documents how he navigated his career to get where he wanted. During the years, he would work on getting an information advantage and then act on that information to be prepared when it matters. He took the time to learn skills that can't be faked and developed a network of peers and friends who sent opportunities his way.

AI ISN'T REPLACING RADIOLOGISTS

Improvements in image interpretation have run far ahead of their adoption. There are hundreds of models that could potentially help radiologists, but AI is often limited to assistive use on a small subset of scans in any given practice. The promise of AI in radiology is overstated by benchmarks alone. Models can lift productivity, but their implementation depends on behavior, institutions, and incentives. So far, radiologists have only become busier as machines have improved.

Transform your workflow →

SPOTIFY TO LABEL AI MUSIC, FILTER SPAM, AND MORE IN AI POLICY CHANGE

Spotify is adopting a new industry standard called DDEX that identifies how AI was used in a song's production.

COREWEAVE EXPANDS OPENAI PARTNERSHIP WITH $6.5B DEAL, BRINGING TOTAL TO $22.4B

This is CoreWeave's 3rd expansion with OpenAI this year.

IN DEFENSE OF AI EVALS, FOR EVERYONE

Evals aren't about following a rigid philosophy - sometimes it's fine to be less rigorous, and sometimes the task demands more rigor.

OPENAI LAUNCHES SHARED PROJECTS FOR CHATGPT BUSINESS

OpenAI has launched shared projects, smarter connectors, and new compliance controls for ChatGPT Business.

US SENATE TO HOLD HEARING ON CRYPTO TAXATION

The Senate Finance Committee has scheduled a hearing on October 1 to review how cryptocurrency should be taxed, reflecting the growing urgency around modernizing the US tax code for digital assets. The session comes as lawmakers debate exemptions for small, routine crypto purchases and broader reforms. Current IRS rules treat all crypto transactions as taxable events, even when no gains or losses occur.

EUROPEAN BANKS JOIN FORCES TO ISSUE MICA-COMPLIANT EURO STABLECOIN

Nine major European banks, including ING, UniCredit, Danske Bank, and CaixaBank, have formed a consortium to launch a MiCA-compliant euro-denominated stablecoin expected to be issued in the second half of 2026. The initiative aims to provide a European alternative to the US-dominated stablecoin market and enhance Europe's strategic autonomy in digital payments, with the new company being established in the Netherlands and seeking e-money institution licensing from the Dutch Central Bank. The stablecoin will offer near-instant, low-cost transactions with 24/7 cross-border payment capabilities, programmable payments, and enhanced supply chain management features leveraging blockchain technology.

POLYGLOBE: A MAP-NATIVE INTERFACE FOR GEOPOLITICS MARKETS

PolyGlobe has introduced a spatial way to browse prediction markets. Traders can now explore elections, conflicts, and macro events via a world map, with narratives clustered by region. An AI engine geotags markets by extracting entities, locations, and coordinates, then organizes them into regional storylines. Upcoming features include map-native search, OSINT overlays (integrating live tweets, flight tracking, and troop movements), and smarter filters, alerts, and watchlists. The goal is to make geopolitics trading more intuitive, visual, and context-rich, moving beyond lists of markets to a dynamic, intelligence-style interface.

WORLD LIBERTY FINANCIAL WILL SOON LAUNCH DEBIT CARD

World Liberty Financial will launch a debit card and retail application "very soon" to integrate its USD1 stablecoin into everyday payments, including Apple Pay integration, co-founder Zak Folkman announced at Korea Blockchain Week 2025. The retail app will merge peer-to-peer transfers with trading features, forming the core of World Liberty's consumer market expansion strategy beyond its current DeFi offerings. The Trump family-backed project has also signed a memorandum of understanding with South Korean exchange Bithumb to explore business opportunities, though specific details of the partnership remain unclear at this time.

UNANSWERED QUESTIONS ABOUT PREDICTION MARKETS

Prediction markets like Polymarket and Kalshi have moved beyond speculation, processing $35B+ since inception and $17.5B in 2025 YTD. Kalshi is leaning heavily into sports (96% of recent volume), while Polymarket skews political and long-tail cultural events. Both face limits from full collateralization and oracle/review structures, yet their growth suggests prediction markets are maturing into one of finance's most speculative and democratizing frontiers.

SOLANA'S RECENT HIGHLIGHTS

Solana ecosystem momentum accelerated with two major Digital Asset Treasury launches: Helius raising $500M backed by Pantera Capital and Solmate raising $300M from UAE-based investors with Ark Invest backing, coinciding with the SEC's approval of generic listing standards, clearing the path for more spot crypto ETFs. Infra milestones included $1 billion of Bitcoin issued on Solana for the first time ever, SP3ND launching to enable Amazon purchases with stablecoins, Orderly Network adding multi-collateral trading support for SOL and USDT, and Reflect Money launching USDC+, a yield-bearing stablecoin. The ecosystem's developer activity surged with the first Solana Mobile Hackathon attracting over 1,000 signups and 276 global submissions.

PEERDAS IS ETHEREUM'S SCALING UNLOCK

Fusaka's headline feature, PeerDAS, lets nodes verify block data by sampling chunks instead of storing full blobs, enabling safer, cheaper L2 data posting today and future L1 scaling. Blob usage just hit six per block for the first time. Counts will rise conservatively at first, then more aggressively as the network hardens. Longer term, even L1 execution data could migrate into blobs, opening more throughput without sacrificing decentralization.

BASE'S NEXT AGE

Base has doubled DeFi TVL to $5B since spring, with Aave and Morpho leading growth (Morpho's BTC-backed loans hit $720M). Tech upgrades like Flashblocks (200ms block times) and Stage 1 decentralization have drawn praise from Vitalik Buterin. A Solana bridge and the new Base App (chat, trade, post, and discover apps) push Base further into mainstream UX, already distributing $500K+ to creators.

EUROPEAN BANKS FORM CONSORTIUM FOR EURO STABLECOIN

Nine major banks, including ING, UniCredit, Danske Bank, CaixaBank, SEB, and Raiffeisen, have launched a consortium to issue a euro-backed stablecoin under MiCA, with a planned rollout in H2 2026.

ALL EVM OPCODES IN ONE DIAGRAM

This technical chart provides a detailed look at every action a smart contract can take on an Ethereum Virtual Machine blockchain, including basic math, transactions, and memory management.

CRYPTO IPO WATCHLIST

Over the next few years, several crypto companies are expected to go public, including BitGo next month, Ripple potentially in 2026, and Consensys in 2027.

BASE APP'S TVL IMPACT

Wasabi Protocol saw a 75% TVL jump when it integrated as a miniapp into the upcoming Base App.

NVIDIA'S POWERFUL AI FACIAL ANIMATION TOOL IS NOW AVAILABLE FOR ALL

Nvidia open-sourced its Audio2Face AI tool, making realistic facial animation technology freely available to all developers. The tool analyzes audio features, such as phonemes and intonation, to automatically generate lip-sync and emotional expressions for digital characters. Audio2Face now includes plugins for Maya and Unreal Engine 5, as well as a training framework for customizing models.