Zuck refuses to give up, secretly develops one of his own called 'Hatch', but the model uses Claude...

marsbitPublished on 2026-08-27Last updated on 2026-08-27

Abstract

Meta, led by Mark Zuckerberg, is developing a new AI agent platform called "Hatch" after its reported unsuccessful attempt to acquire Manus. Described as a "consumer version of OpenClaw," Hatch aims to automate tasks across various digital services. The platform is trained on simulated versions of services like DoorDash, Etsy, Reddit, Yelp, and Outlook, and is deeply integrated with Meta's own ecosystem, including Instagram, Facebook, and WhatsApp. Its goal is to enable end-to-end task completion, such as identifying a product in a social media post, comparing prices, and facilitating purchases. During development, Hatch is reportedly relying on Anthropic's Claude model. However, Meta plans to transition to its proprietary model, Muse Spark, for the official launch, with a newer model named Watermelon expected in October. This phased migration highlights potential technical challenges, as switching core models can affect performance and user experience. Priced at up to $199.99 per month, Hatch targets consumers willing to pay for significant time savings. Its key advantage is default distribution within Meta's vast suite of apps, potentially giving it access to rich user context. However, it faces major hurdles in real-world reliability, needing to handle complex, dynamic web environments, authentication changes, and anti-bot measures that differ from its controlled training simulations. The move represents Meta's strategic push to monetize AI directly beyond its core adver...

Looks like Zuck just can't let it go…

Right after handing Manus back, Meta has turned around and built its own agent platform called "Hatch".

Looks like Zuck just can't let it go…

According to reports from The Information, Hatch functionally emphasizes "setting a goal, automatically breaking it down, and delivering results", which sounds quite similar to Manus. However, Meta insists on a different marketing angle, calling it the "consumer version of OpenClaw".

Its top price tier is $199.99 per month, also matching Manus's highest tier of $200 per month.

Additionally, during its development phase, Hatch used the Claude model. The plan is to switch to Meta's own Muse Spark model only when the product officially launches, with a new model, Watermelon, slated for release in October.

Failed Acquisition? Do It Yourself

Unlike tools like Codex or Claude Code, which are essentially developer tools aimed at writing code, debugging, and building software,

Hatch focuses more on "consuming digital life."

For early training of Hatch, Meta configured environments that included simulated versions of internet services like DoorDash, Etsy, Reddit, Yelp, and Outlook.

This covers local services & delivery, goods trading, community information retrieval, and personal productivity management.

Coupled with Meta's own social network ecosystems of Instagram, Facebook, and WhatsApp,

In other words, Hatch's goal is to bridge the complete path of "Discovery—Comparison—Communication—Scheduling—Purchase."

For example, if a user sees a pair of shoes on Instagram Reels, Hatch could, in theory, identify the product cue, search for similar items, compare prices and delivery times, offer suggestions based on the user's budget and historical preferences, and request confirmation before payment.

However, there are also differences from Manus. Manus is an independent, general-purpose execution environment aimed at research, web automation, coding, and content creation. Hatch, on the other hand, places greater emphasis on long-term memory, personalized context, cross-service operations, and deep integration with Meta's social ecosystem of billions of users.

On the technical roadmap, using Claude during development easily leads others to assume that Muse Spark hasn't yet met the mark in multi-step reasoning, tool calling, and other agent tests.

But long-term dependence on a competitor's model clearly doesn't align with Meta's cost and strategic interests. Once a consumer-grade agent achieves high-frequency usage and long-chain reasoning, inference costs will skyrocket.

Meta's plan is a phased migration.

Use Claude as a transition during the R&D and internal testing phases, switch to the self-developed Muse Spark series of models in the initial productization stage, with subsequent expansion phases taken over by next-generation models like Watermelon, launching in October.

This seems like another potential pitfall—different model behaviors could make migration difficult, right?

However, it's still not confirmed whether Watermelon belongs to the Muse series or will directly become Hatch's core model.

Meta's advertising business has long provided strong cash flow, but the rapid rise of AI infrastructure, model training, and inference costs demands that the company establish more direct AI monetization methods beyond ads.

Hatch's high-end subscription pricing finds its biggest trump card in "default distribution."

Users familiar with Meta's ecosystem don't need to download unfamiliar tools; they can directly use Hatch within Instagram, Facebook, WhatsApp, Messenger, the Meta AI app, or even smart glasses.

A typical challenge for general-purpose Agents is not understanding the user. Meta hopes to use Hatch to convert these scattered signals into the ability to "understand user intent," a capability that pure chat-based AI products lack.

However, a price tag of nearly $200 per month means users will evaluate based on "whether it can consistently save several or even dozens of hours of human effort."

This requires Hatch to maintain reliable performance in real-world scenarios with complex webpages, incomplete information, changing login statuses, payment confirmations, and error handling.

The real internet is far more complex than a simulated training environment. Webpage structures often change, and pop-ups, CAPTCHAs, anti-bot measures, regional differences, and inventory fluctuations can all cause task failures.

This article is from the WeChat public account "QbitAI," author: Meng Chen

Trending Cryptos

Related Questions

QWhat is Meta's new AI agent platform called and how does it position itself?

AMeta's new AI agent platform is called 'Hatch.' It positions itself as the 'consumer version of OpenClaw' and focuses on 'consumer digital life,' aiming to automate tasks like discovery, comparison, communication, scheduling, and purchasing across internet services, with deep integration into Meta's social apps like Instagram, Facebook, and WhatsApp.

QWhich AI model is Hatch using in its development phase, and what are the plans for future models?

ADuring its development phase, Hatch is using Anthropic's Claude model. Meta plans to switch to its in-house model, Muse Spark, for the official product launch. Furthermore, a new model named 'Watermelon' is scheduled for release in October, although its specific role within the Hatch platform is not yet confirmed.

QHow does Hatch differ from other AI coding agents like Codex or Claude Code?

AHatch differs from AI coding agents like Codex or Claude Code, which are developer tools for writing, debugging, and building software. In contrast, Hatch is designed for 'consumer digital life,' automating everyday tasks such as shopping, service booking, and information retrieval across various apps and websites, emphasizing personalization and integration with Meta's social ecosystem.

QWhat is the pricing for Hatch, and what key advantage does Meta have in distributing it?

AHatch's highest pricing tier is $199.99 per month, which is comparable to the top tier of the Manus platform. Meta's key advantage in distributing Hatch is 'default distribution.' It can be directly integrated into and accessed through Meta's massive existing user base on platforms like Instagram, Facebook, WhatsApp, Messenger, Meta AI apps, and smart glasses, eliminating the need for users to download a separate, unfamiliar tool.

QWhat are some potential challenges Hatch might face according to the article?

AAccording to the article, Hatch faces several potential challenges: 1) High subscription cost requires proving it can reliably save users significant time. 2) Real-world internet complexity (changing website structures, pop-ups, CAPTCHAs, anti-bot measures, regional differences, inventory fluctuations) can cause task failures, making reliability difficult. 3) Transitioning from the Claude model to in-house models like Muse Spark and Watermelon may be difficult due to differences in model behavior, posing a potential risk.

Related Reads

US Stock Market Trends (Aug 27): PCE Exceeds Expectations, Pressures Broader Market; Nvidia Rises 4% After-Hours, Nasdaq Futures Up 1%

U.S. Market Trends (Aug 27): PCE Data Weighs on Indices, Nvidia Rises 4% After-Hours, Nasdaq Futures Up 1% U.S. stocks closed slightly lower on Wednesday amid narrow trading. The S&P 500, Nasdaq, and Dow Jones all edged down, ending the Dow's two-day winning streak. The key pressure came from July's PCE inflation data, which showed a 3.7% year-over-year increase, exceeding expectations. While the core PCE met forecasts at 3.3%, the hot headline number boosted Treasury yields and the dollar, dampening hopes for imminent Fed rate cuts. Gold fell below $4,600/oz. Oil prices continued to weaken despite mixed geopolitical signals. The market's real focus was after the close. Nvidia reported Q2 revenue of $96.2 billion, beating estimates, guided for current-quarter revenue to surpass $100 billion for the first time, and projected 70% revenue growth for the next fiscal year. Its shares rose approximately 4% after-hours, lifting Nasdaq futures by about 1%. Amazon's announcement to deploy an additional 2 million GPUs further validated the data center demand narrative for Nvidia. In other sectors, chip stocks like Western Digital and Arm gained. Among the "Magnificent Seven," moves were mixed. Salesforce surged on strong guidance and an expanded partnership with Anthropic. Meta settled a youth addiction case with 29 U.S. states for up to $18 billion, a figure seen as favorable compared to earlier fears. Bitcoin retreated from recent highs, while industrial metals like copper extended their rally. The core market tension remains between inflation concerns pressuring the broader market and strong AI earnings providing sector-specific momentum, setting the tone for early September trading.

marsbit46m ago

US Stock Market Trends (Aug 27): PCE Exceeds Expectations, Pressures Broader Market; Nvidia Rises 4% After-Hours, Nasdaq Futures Up 1%

marsbit46m ago

A Record $70 Billion Inflow in 5 Days! Investors Are No Longer Choosy, Buying Gold and Bitcoin Together

Investors are moving beyond "either/or" choices and are simultaneously pouring money into both gold and bitcoin ETFs. Over the past five trading sessions, ETFs tracking these assets attracted a record $7 billion in combined inflows, pushing some of the largest gold and bitcoin funds to the top of the U.S. weekly ETF inflow rankings. This surge was triggered by U.S. Treasury Secretary's announcement to at least double long-term bond buybacks, which initially pressured Treasury yields and the dollar, boosting prices for both assets. Gold has risen about 13% this month, while bitcoin reclaimed the $80,000 level. The synchronized rally signals the return of the "monetary debasement trade." Amid growing concerns over fiscal sustainability and easing financial conditions, investors are seeking scarce assets perceived as outside direct government control. Gold benefits from its traditional safe-haven role, while bitcoin's fixed supply of 21 million coins positions it as a potential hedge. The SPDR Gold ETF (GLD) attracted nearly $3.4 billion, and the iShares Bitcoin Trust (IBIT) saw $1.5 billion in inflows, both ranking in the weekly top ten. Analysts note the momentum behind the flows is as significant as the volume, indicating investors are aggressively adjusting previously underweight positions. While the narrative of hedging against fiscal stress and currency debasement is gaining traction, some analysts question its sustainability, suggesting equities might be a more reliable hedge in the long run.

华尔街日报51m ago

A Record $70 Billion Inflow in 5 Days! Investors Are No Longer Choosy, Buying Gold and Bitcoin Together

华尔街日报51m ago

Yangtze Memory: Is It the Second ChangXin?

The largest IPO in the history of the STAR Market is approaching. Yangtze Memory Technologies (YMTC) has filed for a listing on the Shanghai Stock Exchange's STAR Market, aiming to raise 33 billion yuan. This surpasses the previous record set by competitor ChangXin Memory Technologies (CXMT), which raised 29.5 billion yuan and saw its market capitalization surge on its debut. Despite both being leading Chinese memory chipmakers founded in 2016 and operating under the IDM model, the two companies are fundamentally different. CXMT focuses on DRAM, the memory used for temporary data processing, while YMTC specializes in NAND Flash, used for long-term data storage. Industry reports indicate the global DRAM market is significantly larger and more concentrated among three major players, where CXMT ranks as the fourth-largest supplier. The NAND Flash market is more fragmented, with YMTC ranking third globally by shipment volume in Q2 2026, though fifth by revenue due to a stronger focus on consumer-grade products. Experts are divided on whether YMTC can replicate CXMT's explosive market debut. Some analysts believe it is highly unlikely, citing a cooler market environment and YMTC's perceived lower strategic scarcity within the AI supply chain compared to DRAM/HBM-focused companies. They warn that aggressive IPO pricing could lead to downward pressure post-listing. Industry forecasts suggest the DRAM market may remain tight, while the NAND Flash market could face price corrections due to new capacity and weaker demand. Other experts argue that while short-term market sentiment differs, the long-term investment value of both companies is comparable, hinging on future performance and potential breakthroughs in areas like HBM. They believe YMTC's IPO is unlikely to face significant cooling given still-elevated market interest in the semiconductor sector.

marsbit1h ago

Yangtze Memory: Is It the Second ChangXin?

marsbit1h ago

Trading

Spot

Hot Articles

How to Buy ONE

Welcome to HTX.com! We've made purchasing Harmony (ONE) simple and convenient. Follow our step-by-step guide to embark on your crypto journey.Step 1: Create Your HTX AccountUse your email or phone number to sign up for a free account on HTX. Experience a hassle-free registration journey and unlock all features.Get My AccountStep 2: Go to Buy Crypto and Choose Your Payment MethodCredit/Debit Card: Use your Visa or Mastercard to buy Harmony (ONE) instantly.Balance: Use funds from your HTX account balance to trade seamlessly.Third Parties: We've added popular payment methods such as Google Pay and Apple Pay to enhance convenience.P2P: Trade directly with other users on HTX.Over-the-Counter (OTC): We offer tailor-made services and competitive exchange rates for traders.Step 3: Store Your Harmony (ONE)After purchasing your Harmony (ONE), store it in your HTX account. Alternatively, you can send it elsewhere via blockchain transfer or use it to trade other cryptocurrencies.Step 4: Trade Harmony (ONE)Easily trade Harmony (ONE) on HTX's spot market. Simply access your account, select your trading pair, execute your trades, and monitor in real-time. We offer a user-friendly experience for both beginners and seasoned traders.

5.3k Total ViewsPublished 2024.03.29Updated 2026.06.02

How to Buy ONE

Discussions

Welcome to the HTX Community. Here, you can stay informed about the latest platform developments and gain access to professional market insights. Users' opinions on the price of ONE (ONE) are presented below.

活动图片