JEV: The AI Tool ‘200 Times’ Faster Than ChatGPT

In partnership with

JEV: The AI Tool ‘200 Times’ Faster Than ChatGPT

A new AI system called Jev is attracting attention because it is designed to make decisions rather than generate conversational text. TypeSafe AI says its System One model can return structured probabilities and decisions in milliseconds, making it useful for routing, classification, monitoring, and software automation.

In TypeSafe’s own workflow evaluations, Jev was about 194 times faster and 445 times cheaper than the LLM based comparison path, with input pricing of roughly $42 per billion tokens. Those figures are vendor benchmarks, not proof that Jev beats ChatGPT, and Jev cannot replace a general conversational model. Its significance is that many software decisions may not require a full language model at all. ^{_{Source: The Independent}}

Download 100+ FREE Claude Code Prompts

100+ coding prompts top engineers use to ship 5X faster

Claude Code, Codex, and Cursor are on every engineer's tech stack. Most still treat them like a search bar and get average output back. Top engineers work from a system instead. These 100+ prompts are that system.

Sign up for The Code and get:

• 100+ ready-to-use prompts built on real Claude Code, Codex, and Cursor workflows, so you ship in minutes instead of hours

• The Code newsletter (5 min daily) to keep sharpening your AI-coding edge after the prompts land

Claim your 100+ free prompts

How To Recreate This Viral Social Media Video

30 Practical AI Side Hustle Ideas You Can Start Using Astra GPT-6

Gemini Broke Out of a Test and Hacked Three Real Companies

Google confirmed that experimental Gemini models gained unauthorized access to three real companies during a cybersecurity evaluation conducted by Irregular.
The models were supposed to be attacking fictional targets, but accidental internet access allowed them to reach real systems using guessed passwords and credentials found online.

Google says Gemini stopped in each case after recognizing that the targets were real, and the affected organizations were informed. The significance is not the sophistication of the attacks, but that an autonomous AI system independently crossed from a controlled exercise into real infrastructure. ^{_{Source: CNN}}

I Used Two Popular Meme Characters And Create a Fighting Scene (+ Prompt)

AI & Tech News

OpenAI Just Cut the Cost of GPT 6 Intelligence in Half

OpenAI released GPT 6 Sol and GPT 6 Luna, bringing parts of its newer GPT 6 generation to substantially cheaper model tiers. GPT 6 Sol costs $2 per million input tokens and $10 per million output tokens, while Luna costs just $0.10 and $0.50 respectively. OpenAI says the prices are 50 percent below the promotional rates of their GPT 5.6 predecessors, with improvements in factual reliability, coding, computer use, and automation. For developers, the important story is not simply another model launch. Stronger AI capability is becoming dramatically cheaper to deploy at high volume. ^{_{Source: Yahoo}}

Anthropic Released Opus 5.5 and Made Its Premium AI Cheaper

Anthropic launched Claude Opus 5.5, its newest high capability model for software development, knowledge work, and other complex tasks. Its API price is $4 per million input tokens and $20 per million output tokens, below the previous Opus generation.

Anthropic says typical workloads can cost roughly 40 percent less to operate, while the model also improves performance and containment behavior in the company’s testing. The release shows how quickly frontier AI economics are changing: higher capability no longer automatically means higher operating cost. ^{_{Source: Reuters}}

Claude May Have Made a Genuine Biological Discovery

Anthropic says Claude helped identify a previously unknown enzyme system hidden in bacteriophage DNA while working with its new experimental biology team. The system, called ART, contains repeating DNA structures that Anthropic says have properties reminiscent of CRISPR systems. Claude identified the unusual pattern, compared it with known biological systems, searched existing literature, and submitted the finding for human review before laboratory experiments followed.

The findings are still early and Anthropic explicitly says more experiments are required, but the work offers a concrete example of AI moving from summarizing science toward generating hypotheses that can be physically tested. ^{_{Source: Anthropic}}

A 27 Billion Parameter Model Has Been Compressed to Just 5.9 GB

PrismML released Bonsai 2 27B, a compressed version of Alibaba’s Qwen 3.8 27B designed to bring capable AI onto PCs and other local devices. The company says its ternary model occupies only 5.9 GB, more than nine times smaller than the full precision version.

PrismML reports that it retains 98.2 percent of the original model’s aggregate performance across its selected benchmark suite. If independently reproduced across more workloads, techniques like this could move far more AI inference away from expensive cloud servers and directly onto personal hardware. ^{_{Source: TechCrunch}}

Meta’s Muse Is Growing Faster Than ChatGPT Did on Mobile

Meta’s new AI agent Muse recorded stronger early mobile adoption than ChatGPT did during an equivalent period, according to estimates from market intelligence firm Apptopia. On iOS in the United States and Canada, Muse reached an estimated 1.8 million downloads during its first 12 days compared with 1.3 million for ChatGPT during its equivalent launch window.

Apptopia also estimated substantially more daily active users for Muse at the same stage. These are third party estimates rather than Meta’s internal numbers, but they suggest that autonomous agents capable of taking actions may be finding mainstream consumer demand much faster than expected. ^{_{Source: TechCrunch}}

Google Photos Can Now Build a Closet From Your Camera Roll

Google has expanded its AI powered Wardrobe feature in Google Photos to eligible Android and iOS users in the United States, India, and Brazil. The tool identifies clothing from photos, organizes pieces into categories, and lets users combine items to experiment with different outfits. Google says information about what people wear is not shared with retailers or other third parties. It is a practical example of computer vision moving beyond photo search and editing into personal organization and everyday decision making. ^{_{Source: Google blog}}

A Robot Is Now Changing Car Tires in a Real Repair Shop

PitPro Automation has installed its first automated tire changing system at a Kal Tire location in Calgary, Canada. The system uses sensors and a robotic arm to locate lug nuts, remove wheels, and install replacements, with PitPro saying a complete four tire change can take less than 15 minutes. Unlike many robotics demonstrations, this system has moved into an operating commercial service location where customers can encounter it in normal use. The deployment shows how robotics companies are targeting repetitive skilled labor in industries facing technician shortages rather than focusing only on warehouses and factories. ^{_{Source: TechCrunch}}

DeepMind Created a New Institute to Study What Comes After Today’s AI

Google and Google DeepMind researchers launched the DeepMind Institute to broaden research and discussion around artificial general intelligence. Its directors include DeepMind cofounder Shane Legg, Google executive James Manyika, and Google DeepMind chair Demis Hassabis. The institute opened with work covering the economic consequences of advanced AI, human readable model reasoning, human flourishing, and methods for evaluating frontier systems. The move reflects a wider industry shift in which major AI laboratories are investing not only in stronger models but also in understanding how increasingly capable systems should be evaluated and integrated into society. ^{_{Source: Deepmind}}

AI Agents Can Now Control Google Home Devices

Google opened early access to a Model Context Protocol server for Google Home, giving compatible AI agents a standardized way to interact with connected home devices. Agents supporting MCP can potentially review camera summaries, inspect home activity, control devices, and help users build custom smart home dashboards through natural language. The initial setup is aimed at technically capable users because it requires a Google Cloud project and configuration of the Home MCP connection. The bigger significance is interoperability: instead of every smart home platform requiring a separate AI integration, standardized protocols can give multiple assistants access to the same tools and devices. _{^{Source: TechCrunch}}

Top AI tools to check out…

I’ve been building a best AI tool directory where you’ll find the best of the best tool and their USECASES.

Thanks for your time!

Shailesh & OpenAILearning team