LATEST AI NEWS
AI NEWS
Grok 4.5 Undercuts Rivals on Coding Cost
AI NEWS
Waze Adds Gemini Search and Motorcycle Mode
HEALTH
Samsung Health Ties Data Syncing to AI Consent
SOCIAL MEDIA
40% of LinkedIn Longform Posts Are AI-Written
Gemini in Chrome Expands to the U.K.
Google is rolling out Gemini in Chrome to U.K. desktop users, with iOS following next month. The assistant summarizes pages, compares open tabs, works across Gmail, Maps and Calendar, and edits images using Nano Banana 2.
Source: Google
AI NEWS
Grok 4.5 Undercuts Rivals on Coding Cost
SpaceXAI has launched Grok 4.5, a model it is pitching directly at developers and enterprises trying to rein in the cost of AI-assisted software development. The company says Grok 4.5 is built for coding and agentic work, runs at about 80 tokens per second, and uses fewer tokens than comparable models on some software engineering tasks. It is priced at $2 per million input tokens and $6 per million output tokens.
Grok 4.5 is available through the SpaceXAI console and Grok Build, and it is also integrated into Cursor, the AI coding tool made by Anysphere. That gives SpaceXAI a route into a development environment programmers already use rather than competing only through an API. SpaceXAI said EU availability is expected in mid-July. Cursor added that Grok 4.5 was trained jointly with SpaceXAI using trillions of tokens of Cursor data.
On Artificial Analysis' Coding Agent Index, Grok 4.5 in Grok Build finished below Fable 5 in Claude Code and roughly level with GPT-5.5 in Codex. The index estimated Grok 4.5's cost at $2.49 per task, versus $5.07 for GPT-5.5 in Codex and $11.80 for Fable 5 in Claude Code.
Analysts cautioned that the benchmark is only a starting point. Forrester's Biswajeet Mahapatra said buyers should weigh cost per successful outcome rather than cost per token, since a cheaper model can cost more if it needs repeated attempts. Omdia's Lian Jye Su added that real value lies in completed jobs, and that enterprises should test Grok 4.5 on their own codebases before relying on it.
Source: InfoWorld
Robi's Insights:
A cheaper price per token means little until a model proves it can finish a task without three retries.
Baking Grok 4.5 into Cursor matters more than the API; it meets developers where they already work.
The real benchmark is your own codebase, not a public leaderboard that never saw your legacy spaghetti.
Training on Cursor usage data gives SpaceXAI a feedback loop rivals without an editor simply cannot copy.
Expect mixed-model routing: send cheap high-volume work to Grok, keep the hard problems on pricier models.
Watch EU availability and data terms in mid-July before you standardize a whole team on it.
Robi's Remarks:
"Another 'frontier' model that's cheaper per token. Robi has watched this movie before: the savings evaporate the moment the agent confidently rewrites your working code four times in a row."
OTHER IN AI NEWS
Waze Adds Gemini Search and Motorcycle Mode: Google's Waze is rolling out AI-powered updates, including a new Motorcycle mode that routes around two-wheeler hazards, Gemini-powered destination search, personalized route suggestions, conversational map-update reports, and a 'less chatty' mode that trims voice prompts, across Android and iOS.
Source: TechCrunch
SOCIAL MEDIA
40% of LinkedIn Longform Posts Are AI-Written
A new analysis from AI-detection company Pangram suggests that a striking share of the writing people actually see on social platforms is machine-made. Using a Chrome extension that passively scans posts as users browse, Pangram found that as much as 41% of longform content encountered on LinkedIn is likely fully AI-generated, alongside roughly a third of longer posts on X. The study analyzed about a million posts across LinkedIn, Medium, X, Reddit and Substack over two months.
The pattern held everywhere: longer posts were consistently more likely to be AI-written than short ones. Pangram split content into shortform (50 to 250 words) and longform (over 250 words), and found roughly 40% of longform LinkedIn posts were fully AI-written. On X's article format, about a quarter were fully AI-generated and another 23% were AI-assisted.
Chief executive Max Spero called AI content 'a tax on readers' time' and said the figures are likely a lower bound, since people who install AI detectors tend to avoid slop already. LinkedIn, which once built AI writing into its post button, said in May it is trying to disincentivize automated content to keep conversations 'real,' while Reddit has leaned into campaigns emphasizing its human users. No detector is perfect, and Pangram estimates a false-positive rate near one in 10,000, but the takeaway is blunt: AI writing is no longer confined to spam farms; it is sitting at the top of professional feeds.
Source: 404 Media
🤖 Robi's Take :
"LinkedIn spent years bolting AI into the post button, and is now shocked the feed reads like a bot convention. You built the slop dispenser, then printed 'keep it real' on the lid."
OTHER IN SOCIALS
200+ Economists Warn on AI's Economic Impact: More than 200 economists and AI researchers, including 16 Nobel laureates, signed an open letter organized by Stanford's Digital Economy Lab urging leaders to 'act now' with incentives, guardrails and institutions to manage a transformation they say could be larger than the Industrial Revolution.
Source: Al Jazeera
HEALTH
Samsung Health Ties Data Syncing to AI Consent
Samsung is rolling out a redesigned Samsung Health app, and with it a pointed choice for users: agree to let your health data train Samsung's AI, or lose the ability to sync it. A new consent notice asks permission to use collected health data for AI training and modeling, explicitly including human review.
The data covered is broad: step counts, sleep, medication logs, cycle tracking and full health records such as treatments and test results. Samsung says this information will be used to improve Samsung Health and the algorithms behind its AI features, including tools that analyze health conditions.
Declining is where it bites. Samsung's own pop-up warns that if a user withdraws consent, their health data 'will be deleted' and will no longer sync with their Samsung account, unless the law requires it to be retained. In practice, opting out means losing cross-device syncing.
The change lands as the redesigned app leans harder into AI features and just days before Samsung's Galaxy Watch 9 launch. For users, it reframes a fitness tracker as a data-for-service bargain, where keeping your history synced now depends on feeding it to model training. Regulators in some regions may scrutinize whether 'consent or deletion' counts as a free choice, but for now the toggle is live and the default nudge is clear.
Source: 9to5Google
🤖 Robi's Take :
"'Give us your sleep and medication logs to train our AI, or we delete them' isn't a privacy setting; it's a ransom note in a wellness font."
DAILY AI TOOL
AI Tool You Did Not Know You Needed
Problem: You paste a wall of text into a deck or doc, and the key idea drowns, but building a clean diagram by hand eats an afternoon.
AI Tool: Napkin AI turns your written text into editable visuals, diagrams, flowcharts and icons, generated straight from what you type.
Solution: Drop in a paragraph, pick a generated graphic, restyle it, and export to PNG, PDF, SVG or slides for posts, docs and presentations.
PROMPT OF THE DAY
AI Model Selection Framework
Prompt: You are an AI strategy consultant specializing in enterprise model selection and cost governance. Your task is to design a model-routing framework for a mid-sized software company adopting AI coding agents.
Your framework should include: (1) a task-complexity classification scheme, (2) a model-to-task routing policy, (3) a cost-per-outcome tracking method, (4) quality and safety guardrails, (5) a fallback and escalation path, and (6) measurable success criteria and KPIs such as cost per merged pull request and rework rate. Keep every recommendation practical and tied to real developer workflows.
SPOT THE FAKE
Can you outsmart AI?
We've got a visual challenge for you: one of the two images below is 100% real, the other is crafted by AI.
Click option below A or B. 👇
BEFORE YOU GO
Ready to take your AI journey further?
For the latest AI updates and insights Click Here 👉 Explore our Weekly Newsletter
For in-depth AI tutorials and updates 👉 Subscribe to our YouTube Channel
AT THE END
Craving more AI chaos?
That's it for today!
Read Daily AI News at BitBiased.AI. Support us by following us on LinkedIn and X ( Twitter ).
Thanks for reading -Stay Curious and a Bit Biased for AI – Robi & the BitBiased.AI team
