@kazani

The BlogChain Newsletterhttps://paragraph.com/@kazani Delivering interesting content on Bitcoin, Security, Privacy, Crypto and AI. It's FREE, Takes less than 5-minutes to read, and you are guaranteed to learn something. Let's front-run the opportunity!Sun, 19 Jul 2026 01:39:37 GMThttps://validator.w3.org/feed/docs/rss2.htmlhttps://github.com/jpmonette/feedenThe BlogChain Newsletterhttps://storage.googleapis.com/papyrus\_images/ba65564709459f0ec415bd2204e42b37.jpghttps://paragraph.com/@kazani

All rights reserved<![CDATA[Samjha Do]]>https://paragraph.com/@kazani/samjha-do cz5N5sDIywEEAOGr8pGeMon, 13 Jul 2026 14:27:59 GMTसमझा दो

Samjha Do means "make me understand." Last week I built an app with that name. It does one thing. You describe a screen that is confusing you, and it explains what is happening in calm Hindi or Marathi, as text and as a voice you can listen to.

I built it for a hackathon run by Mesh API. I also built it for someone in my family who cannot see, and for every parent who has ever called their child in a panic because a message on their phone said something frightening.

If you grew up translating technology for your family, you know the call. "There is a message asking for an OTP. Should I give it?" The answer matters. Sometimes the message is the bank doing its job. Sometimes it is a thief. The whole reason this app exists is that it can tell those two apart. Describe a real one time password request and it explains, calmly, what to do. Describe the scam version and it tells you, gently and clearly, to stop.

What it is, honestly

The app is small. It is one Cloudflare Worker and a single web page. The thinking is done by Mesh API. One model writes the explanation. A second model, a voice model called Bulbul from Sarvam, reads that explanation aloud in Hindi or Marathi. Both sit behind one key.

It took four days. The deadline was seven. By the time I submitted it, the total cost of all the AI work across forty eight test requests was eleven cents.

I am telling you the cost because the loud advice online right now is to build an AI agent and get rich. The part those threads skip is where the real work happens. It is not in the model. It is in everything around it.

Where the work actually was

My laptop is a 2015 MacBook Air with four gigabytes of memory. It cannot run the local test server that a normal build would use. So every time I needed to check my work, I deployed the app to the live internet and tested it there. That kept me honest, because I was always looking at the real thing and not a copy of it.

I attacked my own app before the judges could. I asked it questions it was never meant to answer. Most of the time it refused correctly. But when I asked it something with no screen at all, like a plain trivia question, it invented a fake screen and explained that instead. That is the exact failure that would embarrass the app in front of a real user. I found it, fixed it that evening, deployed the fix, and tested until the invention was gone and nothing else had broken.

I measured the speed instead of hoping. An explanation takes about six seconds. The voice takes about eight seconds to start. I did not pretend those seconds do not exist. I built the demo so that it talks through the wait, which also proves, live, that the app is really calling the service and not faking it.

The moment I made the code public, I realized the way the app talks to Mesh was now public too. Anyone could call it. So I put a hard spending limit on the account. No amount of stranger traffic can run up a bill I did not agree to.

None of that is exciting. All of it is the difference between a demo and something a member of your family can actually use without you standing next to them.

Why this is the niche

The results of the hackathon are not out yet. Mesh API already gave the project a public mention, which meant more to me than I expected. But the app does not need a prize to be useful. It is live right now at samjha-do.kazani.workers.dev. The code is open. The next confusing screen your mother sees has an answer that does not require her to phone you first.

Small agents that do one specific thing, built for the people the industry usually forgets, are where I want to spend my time. Not a general assistant. Not a platform. One clear job, done in a language and a voice the person actually trusts, with the scam caught before it costs them anything.

That is the whole idea. I intend to stay with it.

Share

]]>kazani@newsletter.paragraph.com (Kazani)<![CDATA[BlogChain - July 12 2026]]>https://paragraph.com/@kazani/blogchain-july-12-2026 mdQfEYTts8JnAxMVVaEaSun, 12 Jul 2026 08:08:56 GMT

AI Tools & Agents

Crypto & Web3

Mindset & Ideas

Privacy & Security

Football & Sport

Misc Worth Reading

From Hacker News

That's a wrap. See you next Sunday — kazani

Daily reads on Telegram: Join here

Follow me on Farcaster: kazani

]]>kazani@newsletter.paragraph.com (Kazani)<![CDATA[BlogChain - July 5 2026]]>https://paragraph.com/@kazani/blogchain-july-5-2026 0LGJP3m63MFGmKru6jXcSun, 05 Jul 2026 07:03:45 GMT

AI Tools & Agents

Crypto & Web3

Mindset & Ideas

Privacy & Security

Football & Sport

Misc Worth Reading

From Hacker News

That's a wrap. See you next Sunday — kazani

Daily reads on Telegram: Join here

Follow me on Farcaster: kazani

]]>kazani@newsletter.paragraph.com (Kazani)<![CDATA[The 17.9% Problem: Why LLM Agents Need Real Feedback, Not Self-Verification]]>https://paragraph.com/@kazani/the-179percent-problem-why-llm-agents-need-real-feedback-not-self-verification s9kntnZg5p9ryUaJ9mdqMon, 29 Jun 2026 10:22:01 GMT

The experiment

The researchers built ComPilot — a system that uses an LLM to optimize loops in C programs. The LLM proposes transformation sequences (tile this loop, parallelize that one, interchange these two), a compiler applies them, and the program runs on real hardware. The speedup gets fed back to the LLM, which uses it to decide what to try next.

The result: 3.54x geometric mean speedup over unoptimized code across 150 benchmarks, beating the state-of-the-art polyhedral optimizer (Pluto) on 79% of test cases. Off-the-shelf Gemini Flash. No fine-tuning.

That's impressive but not the interesting part.

The interesting part

To validate the design, they ran an ablation: instead of emitting compiler API commands and getting formal legality checks, have the LLM just rewrite the code directly. Verify correctness by comparing output to the original.

This is what most LLM coding tools implicitly do. Write the code, run the tests, check if it passes.

Result: 14–16% lower speedup, 5.3x more tokens consumed — and 17.9% of "correct" transformations were semantically wrong under random inputs, despite passing the original test.

The code looked right. The tests passed. The program was broken.

This isn't a compiler-specific failure mode. It's a grounding failure mode. If the only feedback signal available to an agent is "does the output look plausible," you will get confident, silent wrongness at a non-trivial rate.

What actually fixed it

Formal legality checking via the compiler's dependence analysis. Not output comparison. Not asking the LLM to double-check. Delegating verification to something that can actually verify.

And then: measuring real execution time and feeding that number back.

Two separate signals, both essential:

The LLM's job is exploration and strategy. The environment's job is ground truth. Mix those up and you get the 17.9% problem.

The feedback loop ablation

They also tested removing feedback entirely — let the LLM propose transformations but don't tell it what happened. Blind search.

Single-run speedup dropped 23%. With GPT-4o it dropped 40%.

This is the part worth sitting with. The performance difference between "LLM with feedback" and "LLM without feedback" is larger than the difference between any two LLMs in their comparison. GPT-4o with feedback beats GPT-o3-mini without it.

The model choice matters less than whether the model can observe the consequences of its actions.

What this means for agent design

Three things I take from this:

1. Self-verification is load-bearing in ways that break silently. If your agent validates its own output — by re-reading it, running it against a single test, checking if it looks right — you have a 17.9% silent failure rate floor, minimum. The exact number will vary by domain, but the failure mode is structural.

2. Delegate verification to the environment, not the model. Wherever there's a formal check available — type system, schema validation, compiler, database constraint, test suite with random inputs — use it. Don't ask the LLM to reason about correctness. Route the action through something that actually enforces it.

3. The feedback signal quality is the ceiling on agent performance. ComPilot's feedback is unusually clean: binary legality + continuous speedup ratio, measured on real hardware. Most agents have murkier feedback — user ratings, task completion proxies, soft metrics. The messier the feedback, the harder it is for the agent to learn from it in-context. Designing clear feedback signals is underrated agent infrastructure work.

The meta-point

The paper's framing is compiler optimization. The actual contribution is an empirical demonstration that feedback-grounded agentic loops work, and a controlled ablation showing exactly how much each component contributes.

Most agent papers either show a capability ("the LLM can do X") or benchmark a prompt strategy. This one answers "what does the loop buy you" and "what breaks if you remove formal verification" with actual numbers.

That's rare. The numbers are:

If you're building anything where agents take actions with real consequences, those numbers are the architecture argument you've been looking for.

https://arxiv.org/pdf/2511.00592

Share

]]>kazani@newsletter.paragraph.com (Kazani)llmaiagents<![CDATA[BlogChain - June 28 2026]]>https://paragraph.com/@kazani/blogchain-june-28-2026 aek7F3YyuDoougpy0MJeSun, 28 Jun 2026 08:16:53 GMT

AI Tools & Agents

Mindset & Ideas

Privacy & Security

Football & Sport

Misc Worth Reading

That's a wrap. See you next Sunday — kazani

Daily reads on Telegram: Join here

Share

]]>kazani@newsletter.paragraph.com (Kazani)<![CDATA[Bodies together, Minds elsewhere.]]>https://paragraph.com/@kazani/bodies-together-minds-elsewhere QE0O4UKxgHsCUXCZDvYtSat, 10 Jan 2026 13:26:40 GMT

How screens have changed in-person connection?

Screens let us edit, delay, and curate interaction. That trains the nervous system to expect control.

Face-to-face interaction is the opposite: real-time, uneditable, full of micro-risk (tone, pauses, eye contact).

So when people say “in-person feels harder now,” it’s often because:

This isn’t weakness - it’s conditioning. We haven’t lost the capacity for deep in-person connection. We’ve lost conditioning

Attention fragmentation leaks into presence

Screens train:

In person, presence requires:

So even when two people are physically together, their attention habits may still be optimized for screens, not humans.

This creates a mismatch:

Bodies together, minds elsewhere.

Just like physical strength:

Awkwardness is not a flaw. It’s a signal of underused capacity.

We notice:

We miss:

So the story isn’t:

“Screens ruined connection.”

It’s closer to:

“Screens raised the contrast between shallow and real connection.”

Screens didn’t destroy our ability to connect.

They exposed how much connection depends on trained attention and emotional tolerance.

In other words:

The people who rebuild:

often report deeper in-person connection than pre-screen life, because it’s now chosen, not accidental.

Face-to-face connection has always required capacity:

Screens didn’t remove those requirements.

They just gave us a way to avoid training for them.

Avoidance feels easier, until you try to return.

]]>kazani@newsletter.paragraph.com (Kazani)wordsscreens<![CDATA[Now that Base App is open for all, we're onboarding everyone to Base]]>https://paragraph.com/@kazani/now-that-base-app-is-open-for-all-were-onboarding-everyone-to-base VqtYj63WaYVrSW4XbcnaSun, 21 Dec 2025 11:32:40 GMTTLDR: The new Base App is available in more than 140 countries. It’s an everything app for social, trading, and payments with countless ways to earn. Download it today to discover the internet’s best assets, earn from your content, and trade instantly from the new social feed.

[Get the Base app \ \ Join Friends to trade, earn and discover\ \ https://base.app\\ \

I think these launches are genuinely important for the industry, especially coming from major players like Coinbase.

The Base App - and anything tied to the Base chain that eventually leads to a Base token - feels like one of the clearest opportunities for next year, more so than most alternatives.

There's a real fight to pull crypto-native users into a new social layer.

With Farcaster stepping back from that race, Base App suddenly has a strong opening, largely because of the resources behind it. That's why being more active there during a bear market actually makes sense. I'm planning to start posting on Base App next year, especially since I've heard the Base airdrop will heavily weight Base App usage.

They're also running a Holiday Rewards program that's flying under the radar. It's live until December 21, and participating now likely puts you in the early-user bucket. Inviting others could matter a lot for potential rewards.

The steps are simple:

- Connect your social account to Base App

- Make a trade of $1 or more and keep at least $5 on balance

- Invite friends who do the same

INVITE LINK: https://base.app/invite/friends/3VZ5JQ97

Each completed step earns points, which translate into a share of a USDC rewards pool of up to $2 million.

[Get the Base app \ \ Join Friends to trade, earn and discover\ \ https://base.app\\ \

Download the BASE App TODAY

]]>kazani@newsletter.paragraph.com (Kazani)baseairdropbasepostingfarcaster<![CDATA[Privacy doesn't mean anything anymore, anonymity does]]>https://paragraph.com/@kazani/privacy-doesnt-mean-anything-anymore-anonymity-does xQ2QYegXYaZhsqj90yh3Sat, 20 Dec 2025 11:36:21 GMT

1. "Privacy" has been hollowed out

In modern tech, privacy is mostly a policy promise, not a technical constraint.

If a system:

- collects identifiers

- stores metadata

- logs access

- can be compelled to disclose

Then "privacy" is conditional trust, not protection.

Most "privacy-first" products are doing risk management, not risk elimination.

That's not cynicism; that's architecture.

2. Anonymity is architectural, not moral

True anonymity means:

- the system cannot know

- therefore it cannot leak

- therefore it cannot comply

This is the only model that survives coercion.

If a service could identify you under pressure, it eventually will. Law, breach, insider threat, or acquisition - pick one.

3. The "Privacy Theater Playbook" describes how services collect user data incrementally through email registration, password resets, phone verification, and identity confirmation, despite privacy policy claims.

Incremental identity accretion is exactly how most systems launder surveillance while maintaining plausible deniability:

"Just an email"

"Just for recovery"

"Just for abuse prevention"

"Just for compliance"

Each step seems reasonable in isolation.

In aggregate, it recreates full-spectrum identity.

That's not accidental. It's incentive-aligned.

4. Possession = vulnerability

This is the core law most people refuse to internalize:

"If you hold the data, you are the risk surface."

No amount of encryption, policy language, or goodwill overrides that.

The safest data is data that never existed.

Privacy still matters - just not as a primary defense.

Privacy today means:

- data minimization

- compartmentalization

- harm reduction

It's a damage control layer, not a shield.

The term "privacy" in tech is often misused as a marketing tactic rather than an architectural reality, with many services collecting extensive user data while claiming to protect privacy.

True anonymity is an architectural decision that makes it technically impossible to compromise or identify users, even under duress or legal orders.

The core vulnerability in most services is possession of user data; services that do not collect or store certain data cannot leak or be forced to reveal it.

Mullvad VPN exemplifies real anonymity by using randomly generated account numbers instead of personal information, rendering them unable to comply with data requests.

Email addresses are identified as a primary mechanism that destroys anonymity due to their role as identity markers, trackability, persistence, and susceptibility to social engineering.

Crypto payments are accepted to decouple transactions from persistent identity, as traditional payment systems create surveillance infrastructure, though traditional payment options are also pragmatically supported.

5. Crypto ≠ anonymity by default

- Public blockchains are anti-anonymity by design

- Pseudonymity collapses under correlation

- Crypto payments only help if the entire stack respects anonymity

Otherwise crypto becomes the most permanent surveillance ledger ever created.

Anonymity is distinguished from impunity, security, invisibility, and zero trust, emphasizing that it's about architectural limitations on data collection to minimize damage when trust fails, not a license for illegal activity or complete invulnerability.

Privacy without anonymity is a promise.

Anonymity is a constraint.

Most users fail not because they misunderstand anonymity, but because they achieve it in one layer and leak it in another.

Email is one such leak - but not the only one:

- timing correlations

- writing style

- device fingerprinting

- payment rails

- social graph leakage

Anonymity is about not creating irreversible power asymmetry.

It's a defensive posture against:

- future regime change

- policy drift

- data reuse

- retroactive enforcement

- adversaries you haven't met yet

That temporal dimension is often missed.

6. Most products will never choose anonymity by default because:

- it breaks growth analytics

- it breaks personalization

- it breaks monetization

- it breaks compliance narratives

- it breaks VC expectations

Which means:

"Anonymity will not be mass-adopted. It will be selectively adopted by people who understand power."

That"s the real dividing line.

]]>kazani@newsletter.paragraph.com (Kazani)privacyanonymity<![CDATA["Where are you supposed to go if you don't care about growth?"]]>https://paragraph.com/@kazani/where-are-you-supposed-to-go-if-you-dont-care-about-growth 0xwTc683Qxx2mK7Hnc1QTue, 09 Dec 2025 13:56:24 GMT

If you truly don’t care about growth, there is only one “destination” available to you:

You don’t go anywhere.

You stay exactly where you are - until life forces you to move.

And that’s not a poetic statement. It’s a structural truth of the HUMAN VECTOR system.

Growth isn’t a lifestyle preference; it’s how reality prevents collapse.

1. “Not caring about growth” is itself a developmental signal

It tells me you’re not actually asking about growth - you’re asking about relief.

People say “I don’t care about growth” when:

This is not apathy.

It’s compression - your system is overloaded.

Your problem isn’t lack of ambition.

Your problem is unsustainable conditions.

2. If you don’t intentionally grow, you get pulled backward

Human development is not stable by default.

In the HUMAN VECTOR model, every quadrant has entropy:

Stagnation is not neutral.

Stagnation leads to regression.

Regression leads to crisis. Crisis forces growth anyway - only violently.

So the truthful answer is:

If you don’t choose a direction, life chooses it for you.

And it usually chooses the worst possible timing.

3. You don’t need “growth.” You need integration.

You might imagine growth as:

But growth in this model means something much simpler:

👉 solving the specific problem that is currently limiting your experience of life.

If you don’t care about “growth,” fine. Then ask one question:

What problem is hurting you the most right now?

Solve that, not “growth.”

Because development is just problem-solving:

“The quality of your life is determined by the quality of the problems you’re solving.”

You don’t need purpose.

You don’t need optimization.

You don’t need transformation.

You need a next problem small enough to solve and meaningful enough to matter.

4. If you feel like you don’t care, that’s a Spirit–Mind collapse

From the framework’s perspective:

When these two drop simultaneously, the system defaults to:

“Why bother?”

But that’s not truth.

That’s a signal that one quadrant has become your constraint.

When both go dark, you lose the felt sense of direction.

The fix is not motivation.

The fix is reconnecting to one small source of aliveness - not growth.

5. So where do you go?

There are only three real options:

Option A - You go deeper into your current state (regression).

Life becomes narrower, smaller, more automatic. Eventually something breaks. You’re pushed into growth through pain.

This path is common and unnecessary.

Option B - You pick the smallest quadrant with the highest leverage.

Not “growth.”

Not “purpose.”

Not a five-year plan.

Just the minimum effective dose that reduces suffering:

Tiny actions solve big problems because quadrants cascade.

Option C - You redefine the game entirely.

If growth feels meaningless, consider the possibility that:

You’re not done growing.

You’re done growing in the old way.

This is classic Phase X.1 → X.2 transition:

Old identity exhausted.

New identity not yet formed.

The correct move here is not growth.

It’s uncertainty navigation.

You step into a period where you don’t know who you are or where you’re going - and you stop trying to force an answer.

You let the new interest cycle reveal itself.

Growth returns when the right problem appears, not when you demand motivation.

6. The real question underneath "the" question

Question - if I strip away the phrasing - is not:

“Where do I go if I don’t care about growth?”

It’s:

“Is there a place for me in this model if I’m exhausted, directionless, or just done trying?”

Yes.

Absolutely.

The model expects this phase.

It calls it:

Phase X.1 – Dissonance

The exact state before real transformation begins.

This isn’t the end.

This is the precondition.

[ChatGPT - Vertex Chatbot - Human Vector \ \ ChatGPT is your AI chatbot for everyday use. Chat with the most advanced AI to explore ideas, solve problems, and learn faster.\ \ https://chatgpt.com\\ \

]]>kazani@newsletter.paragraph.com (Kazani)growthwords<![CDATA[Time is money. Bitcoin is time.]]>https://paragraph.com/@kazani/time-is-money-bitcoin-is-time ayQg8CEfLFdISWNdDqwXSun, 07 Dec 2025 11:06:48 GMT

This line sounds poetic, but there's a hard structural truth underneath it.

1. Money = Stored Time

Every unit of money represents:

Money is just a ledger for human energy.

2. Fiat breaks the time–value link

When central banks inflate the supply, they're effectively diluting the stored time of everyone who earned honestly.

Your past hours become worth less.

Your future hours become more expensive to maintain the same life.

Inflation is a tax on time.

3. Bitcoin re-attaches money to time

Because Bitcoin has:

It turns money back into a scarce container for human time.

Your effort today is preserved tomorrow.

No one can reach into your stored time and dilute it.

4. Energy-backed, not trust-backed

Bitcoin converts energy + time + computation into monetary units through proof-of-work.

That’s why Bitcoiners say it's "time stamped in blocks".

Mining literally embeds the cost of time into the asset.

5. Why this idea resonates

People feel that modern money is stealing from them, even if they can't articulate the mechanics. Bitcoin flips the model:

If time is the fundamental human resource, then the fairest money is the one hardest to inflate.

Bitcoin is that system.

]]>kazani@newsletter.paragraph.com (Kazani)bitcoinmoneytime<![CDATA[Crypto doesn't care about privacy because incentives reward visibilty]]>https://paragraph.com/@kazani/crypto-doesnt-care-about-privacy-because-incentives-reward-visibilty jXfz0NdJsWsI6elYyPLtTue, 02 Dec 2025 12:11:44 GMT

This is the uncomfortable truth most people in crypto refuse to confront:

We say we care about privacy, but we keep building on architectures that structurally cannot provide it.

Not because we don't know better, but because the incentives reward visibility, speculation, and network effects, not privacy.

Let's cut through the illusions.

1. Public blockchains were never designed for privacy

They were designed for:

Privacy was not a design goal. At best, it was an afterthought masked by pseudonymity.

If you build on a system optimised for transparency, you inherit transparency.

2. Builders chase liquidity, users chase convenience

You can't separate the psychology from the architecture.

Privacy tech requires friction: new tools, new UX, new patterns, new trust assumptions. Most users won't touch that unless privacy is forced upon them.

3. Public state is a feature for founders but a surveillance nightmare for users

Founders love:

But global state is a surveillance oracle. Every wallet, every action, every relationship becomes a permanent data trail.

4. Privacy cannot be bolted on

Privacy must be the default or it will never materially exist.

What happens when privacy is optional?

Optional privacy = no privacy.

5. The real reason no one builds privacy by default

Privacy hurts the narratives that built crypto:

It also makes compliance harder. And VC due diligence harder. And user acquisition tracking harder.

Most of the industry is not ideologically aligned with privacy; it is ideologically aligned with appreciation, liquidity flow, and public metrics.

6. We keep pretending "decentralized = private"

It isn't. Never was. Never will be, without deliberate cryptographic design.

Transparency is not freedom. It's simply a different kind of surveillance.

If we truly cared about privacy, we would:

7. The problem is not "no privacy tech"

The problem is no incentive to adopt it.

Until:

...nothing will change.

Privacy won't emerge until builders refuse to keep enabling surveillance-by-default architectures.

And until users stop treating transparency as " neutral"

]]>kazani@newsletter.paragraph.com (Kazani)privacy<![CDATA[Soft De-Google Guide (no rooted device nor custom ROM required)]]>https://paragraph.com/@kazani/soft-de-google-guide-no-rooted-device-nor-custom-rom-required J0y0cMjY0iS6ca4leTfwTue, 25 Nov 2025 11:43:06 GMTSoft degoogling doesn’t require a custom ROM, but you can still cut your Google exposure by most of the way. You keep your stock Android setup while stripping out Google’s bloat, replacing their apps, tightening permissions, and routing all connections through privacy-respecting tools.

Start with Shizuku, since it gives certain apps elevated permissions without rooting. Install the APK, enable Wireless Debugging, pair it, set its battery mode to unrestricted, and get it running.

Then replace Google Play with F-Droid (use Droid-ify), install App Inspector, and use aShell (or any ADB shell tool) with Shizuku permissions. Identify every Google package you want gone and uninstall it with

pm uninstall -k --user 0 package-name.

If you remove system apps like Messages, Keyboard, or Dialer, make sure you’ve already installed alternatives and granted permissions. You can reinstall any package with

pm install-existing package-name.

For alternatives, FOSS apps cover most essentials:

Fossify for basics, Quick or Right Messages for SMS, Ente for photos, Proton services for mail/storage, DuckDuckGo or other privacy browsers, HeliBoard or Fossify Keyboard, Notesnook or Standard Notes for notes, Signal (use Molly - Signal Fork) through APK, ReVanced for patched media apps, and OpenStreetMap/OsmAnd/Organic Maps for navigation.

Add tracking protection through TrackerControl, DuckDuckGo’s protection, NetGuard, AdGuard, or similar tools. These act as a private DNS layer and let you manage all outgoing connections app by app. On-device Private DNS should be off if you use these, otherwise configure a trusted DNS like ControlD, Quad9, Mullvad, or AdGuard.

Freezing apps with SuperFreezZ or Heil keeps them fully shut down. Grant any needed permissions through ADB commands if they aren’t available in settings.

Update system privacy settings:

Delete your advertising ID, restrict location access, deny mic/camera by default, pause Google’s activity tracking, and turn off diagnostics. Enable app hibernation for anything unused. Consider a separate work profile via Shelter for apps you don’t fully trust.

It’s an ongoing process. Keep refining your setup and, when possible, switch to a non-Google OS for an even cleaner slate.

Recommend: Tools like Canta or UAD can handle package removal through a GUI if you don’t want to run commands manually.

]]>kazani@newsletter.paragraph.com (Kazani)privacy<![CDATA[Unmasking Surveillance]]>https://paragraph.com/@kazani/unmasking-surveillance wHp9xNX9kWqYaKviVIndThu, 13 Nov 2025 11:47:13 GMT

If you're serious about #UnmaskingSurveillance, start by understanding the full stack, not just government spying, but corporate telemetry, algorithmic profiling, and self-inflicted data leakage.

Most people fear "surveillance" but still run full-telemetry OSes, sync their thoughts to cloud AI, and carry always-on microphones. Hypocrisy, not ignorance, sustains the system.

1. The Reality: Surveillance Is Layered

A. State layer:

– Agencies (NSA, GCHQ, etc.) operate through partnerships with telecoms and cloud providers.

– Metadata, not content is the gold mine: who you talk to, when, where, how often.

– Legal cover comes from "national security" frameworks and secret courts, not warrants.

B. Corporate layer:

– Google, Meta, Microsoft, Amazon, TikTok don't "spy", they monetize prediction.

– Every touch (scroll, dwell time, cursor hover) trains behavioral models to forecast what you'll buy, believe, and click.

– They've built better psychological dossiers than any intelligence service ever could.

C. Social layer (self-surveillance):

– The illusion of voluntary exposure, location sharing, social media stories, smart homes makes you your own informant.

– Every digital convenience is a trade: speed for sovereignty.

2. The Lie You Tell Yourself

" I have nothing to hide."

False.

You have everything to protect, your future optionality, your freedom to dissent, your ability to make choices without algorithmic manipulation. Surveillance doesn't just watch you; it shapes you. The goal isn't data, it's compliance.

3. The System's Leverage Points

– Data concentration: Every app that centralizes identity becomes a surveillance hub.

– Default settings: 90% of tracking persists because people never alter defaults.

– Network effects: Platforms weaponize "everyone's here" to eliminate opt-out feasibility.

– AI intermediaries: The next phase of surveillance is not human observation, but model-based inference. You'll never see the watcher.

4. If You Actually Want to Unmask It

1. Decentralize your identity, use self-hosted tools or privacy-first providers.

2. Run de-Googled OS variants (GrapheneOS, LineageOS, Linux).

3. Replace convenience apps with privacy counterparts:

4. Block data egress: firewall telemetry, use Pi-hole, Private DNS like NextDNS or ControlD, Tailscale-based routing.

5. Cut dependency chains: Don't just switch browsers; stop syncing everything through one cloud.

6. Assume inference, not just observation: even anonymized data trains models about you. Your pattern is your identity.

5. The Deeper Shift

Stop thinking of privacy as secrecy, it's self-possession.

Surveillance thrives on your addiction to convenience and validation.

To unmask it, you first have to unmask your own complacency.

You can't fully escape the panopticon, but you can refuse to feed it.

Surveillance loses power when you stop being predictable.

Map your digital footprint, every account, every sync, every API dependency.

Then systematically dismantle or replace one per week.

You don't need slogans. You need discipline.

]]>kazani@newsletter.paragraph.com (Kazani)privacysurveillance<![CDATA[Offline Privacy Awareness Gap]]>https://paragraph.com/@kazani/offline-privacy-awareness-gap aIWaY1el8bWGQpbcakqiSat, 08 Nov 2025 12:08:42 GMT

A shift in public sentiment to prioritize offline privacy with the same urgency as online privacy is contingent upon a confluence of factors rather than a single timeline. The current landscape reveals a significant gap in perception and concern, driven by psychological biases, powerful commercial and governmental incentives for surveillance, and a media narrative that often frames offline tracking as a public safety imperative. However, emerging legislative battles, organized public opposition, and the rapid convergence of surveillance technologies suggest that a tipping point could be reached when the abstract risks of offline tracking translate into tangible, widespread harm or when a catalyzing event exposes the systemic dangers of an integrated surveillance infrastructure.

The Perception Gap: Online vs. Offline Surveillance

Public opinion data indicates a clear disparity in how Americans perceive the prevalence and threat of online versus offline tracking, which forms the foundation of the current apathy.

Perceived Prevalence of Tracking

Americans are more likely to believe their online activities are being monitored than their offline movements.

Stated Concerns and Perceived Control

Despite the lower perceived prevalence of offline tracking, Americans consider their physical location highly sensitive information, second only to their Social Security number. An overwhelming majority feels they lack control over data collected by both companies (81%) and the government (84%). However, the feeling of having no control is slightly less pronounced for physical location compared to online data like search terms (41%) and websites visited (48%). This suggests that while offline data is valued, the threat feels less immediate or comprehensive.

Support for Visible Surveillance

Public support for visible, government-operated surveillance remains high when framed around security.

Psychological and Structural Barriers to Concern

Several factors create inertia, preventing stated privacy concerns from translating into widespread public action against offline surveillance.

The Privacy Paradox and Calculus

The "privacy paradox" describes the disconnect between people's stated desire for privacy and their actual behavior, where they readily divulge personal information. This is particularly acute with offline surveillance technologies that offer tangible benefits.

Habituation and Surveillance Creep

The constant visibility of cameras in public spaces leads to psychological desensitization.

The Proliferation of Private and Public Surveillance Networks

A powerful ecosystem of corporate and government interests is driving the rapid expansion of offline surveillance, often outpacing public awareness and regulatory oversight.

The "Safety-as-a-Service" Business Model

Companies like Flock Safety have developed highly scalable and lucrative business models built on widespread data collection.

The Role of Government Funding

Federal grant programs are a primary catalyst for the adoption of surveillance technology by local law enforcement, effectively subsidizing the market for private vendors.

Nascent Resistance and Emerging Policy Debates

Despite these barriers, there are clear signs of growing concern and organized pushback against certain forms of offline surveillance, indicating that a shift is already underway in specific domains.

Legislative and Regulatory Action

Public Opposition and Corporate Accountability

Catalysts for a Potential Tipping Point

A significant shift in public sentiment toward offline privacy will likely require one or more catalyzing developments that make the abstract threat of surveillance concrete and personal.

1. High-Profile Misuse and "Offline Cambridge Analytica"

Historical precedent shows that scandals are powerful catalysts. The Edward Snowden leaks dramatically shifted public opinion on government surveillance, with concern that anti-terror policies went too far jumping from 35% to 47%. Similarly, the Cambridge Analytica scandal took corporate data privacy concerns mainstream. A similar event involving offline data—such as the documented use of Flock's network to search for a woman who had a self-administered abortion or for thousands of ICE-related searches—could serve as a tipping point if it receives widespread media attention and demonstrates direct harm to ordinary citizens.

2. The Convergence of Surveillance Technologies

The integration of previously separate surveillance systems into a single, powerful network represents a qualitative shift in tracking capabilities that could trigger public alarm.

3. A Shift in Regulatory Environment and Media Framing

The current U.S. approach to privacy is fragmented compared to the EU's comprehensive GDPR, which mandates principles like "Privacy by Design". A major federal privacy law in the U.S. could fundamentally alter corporate behavior and public expectations. This is often coupled with media narratives. Currently, local news often frames surveillance tools as crime-fighting successes. A sustained shift in media focus from anecdotal crime-solving stories to the systemic risks of mass surveillance, wrongful arrests due to algorithmic error, and data misuse could erode the public's acceptance of the security-privacy trade-off.

4. Increased Knowledge and Transparency

There is a direct correlation between knowledge of surveillance practices and public concern. One study found that individuals with more knowledge of ALPR usage had "significantly lower levels of trust in police"ORS: Documents-Research Briefs-ALPR (In Detail) | Division of Criminal Justice. As advocacy groups and journalists continue to expose the scale of data collection, the opacity of private-public partnerships, and the flow of data to federal agencies, public awareness will grow, potentially eroding the foundations of the privacy paradox and leading to greater demand for accountability and regulation.

]]>kazani@newsletter.paragraph.com (Kazani)privacy<![CDATA[Graphene OS: an Android version with enhanced security features]]>https://paragraph.com/@kazani/graphene-os-an-android-version-with-enhanced-security-features ZxZYvX5jXodKTfoN4mDJSun, 27 Jul 2025 11:07:44 GMTPeople often place significant trust in their phones, which have access to a vast amount of personal and sensitive information — such as our locations, financial details, and communications. As a result, phones, even from less prominent individuals, can become valuable targets. Android devices operate on some level of free software, which ideally should allow them to serve their owners' interests. However, standard Android installations generally do not fully meet this expectation. The GrapheneOS Android rebuild is an attempt to improve on that situation.

[GrapheneOS: the private and secure mobile OS \ \ GrapheneOS is a security and privacy focused mobile OS with Android app compatibility.\ \ https://grapheneos.org\\ \

GrapheneOS originally began as "CopperheadOS," a project reviewed here in 2016. However, a few years later, a serious disagreement between the project's two founders resulted in its collapse. Daniel Micay, one of the founders, carried on with the project and developed what is now known as GrapheneOS. According to its history page, GrapheneOS is an independent, open-source initiative that " will never again be closely linked to any specific sponsor or company."

A Canada-based foundation established in 2023 supports the work on GrapheneOS, but there seems to be very little public information about this organization.

Essentially, GrapheneOS aims to enhance Android's security against various threats and to prioritize the privacy of its users. It is built on the Android Open Source Project but eliminates significant code and incorporates numerous modifications. Some of these changes, like a fortified malloc() library or extra control-flow-integrity features, are largely unnoticed by users (unless they cause apps to malfunction, which has reportedly occurred). Some are clearer, yet it's evident that considerable effort has been made to ensure the security enhancements are as inconspicuous as possible.

Installation

Certain Android variations focus on supporting a broad array of devices, aiming to extend the functionality of older models. GrapheneOS is not among these initiatives. It offers support for a limited selection of devices, specifically the Google Pixel 6 through Pixel 9 series, with some minimal support for Pixel 4 and 5 models. However, newer devices are highly recommended.

The 8th and 9th generation Pixels offer at least 7 years of support from launch, increasing from the previous 5-year minimum. Additionally, these models include support for the highly powerful memory tagging security feature, thanks to the adoption of new ARMv9 CPU cores. GrapheneOS employs hardware memory tagging by default to safeguard the core OS and user-installed apps that are known to be compatible from being exploited. Users have the option to apply this protection to all apps, while allowing them to opt-out individually for apps that are not compatible.

My phone had been suggesting for some time that it wouldn't be reliable in the future, but the thought of purchasing a new one filled me with dread. Every new model seems to include more privacy-invading "features" and intrusive AI "assistants," and locating all the "disable" options is a time-consuming and error-prone process. This, combined with the news that Google's "Gemini" appears to have growing access to a device owner's data no matter its settings, motivated the acquisition of a Pixel 9 device to experiment with GrapheneOS and see if it could serve as a daily alternative to the default Android.

Installing firmware on a costly device can be nerve-wracking; the GrapheneOS installer aims to reduce the anxiety involved. The documentation outlines two installation methods: one using a web-based interface and the other via the command line. Of course, I opted for the command-line version. The steps are simple: download the installation image, connect the device, and execute the provided script. The script completed and confidently announced success, yet the device still only booted into standard Android—a consistent outcome, but not the desired one.

[GrapheneOS installation \ \ Installation instructions for GrapheneOS, a security and privacy focused mobile OS with Android app compatibility.\ \ https://grapheneos.org\\ \

After some research, it was discovered that the web installation method is considered more reliable than the command-line version, though this wasn't documented. I gave it a try, and it worked perfectly, marking the start of the GrapheneOS experiment.

Stock Android offers convenient features to simplify transitioning to a new device, which is not surprising considering the motivation to encourage frequent upgrades. Most of the data, apps, and settings from the previous device are automatically transferred to the new one. In contrast, GrapheneOS lacks this capability; a newly set up phone is a blank canvas that requires configuration from scratch. You can anticipate spending considerable time rediscovering those settings that were perfectly adjusted some years back.

A stock Android installation includes a wide array of apps from the start, many of which the user probably didn't want and often can't remove. GrapheneOS lacks all these unnecessary apps. It provides its own versions of a web browser, camera app, PDF viewer, and app store. Notably, GrapheneOS doesn't come with the Google Play Store or any of its apps (though keep reading for more on Google Play). The app store only has a total of 13 apps.

The web browser, named Vanadium, is a Chromium fork. It allows strict site isolation on mobile devices, a feature apparently lacking in Chrome, and includes several code-hardening features. The documentation strongly advises against using Firefox, labeling it as " more vulnerable to exploitation."

The camera app is claimed to be the finest in a style commonly associated with GrapheneOS:

GrapheneOS Camera surpasses all portable open-source camera options and even most proprietary camera apps, including paid ones. On Pixels, the Pixel Camera can serve as an alternative offering more features.

The camera app strips Exif metadata by default, and location metadata must be enabled separately if it is wanted.

App stores

Another option available from the GrapheneOS store is the Accrescent app store, an alternative repository emphasizing security and privacy. It offers a selection of additional apps, such as Organic Maps, the Molly Signal fork, and IronFox, a fortified version of Firefox.

With those app stores, you can activate some basic phone functions, but unfortunately, many of us require a little more than just that. One option, however, is F-Droid, which can be installed and used on GrapheneOS. Although those deeply focused on security, like members of the GrapheneOS community, often criticize F-Droid (as illustrated in this article), it remains a valuable resource for mainly free software apps.

Ultimately, many people frequently rely on the Google Play store; without the apps available there, an Android device can be almost unusable for some. GrapheneOS provides a sandboxed version of Google Play, making it just a regular app without the special privileges it usually holds on standard Android systems. It functioned perfectly in this instance; although the documentation notes that some apps might not function properly, I didn't experience any issues.

It is important to mention that Android offers an " integrity API" which allows checking the status of the software operating on the device. It can confirm, among other things, whether the secure-boot sequence was successfully completed or if the device is using an official Android version. GrapheneOS incorporates this API and, because it utilizes the secure-boot mechanism, it can clear the initial test, but it is not an authorized image and fails the second test. Some applications are concerned with the outcomes of these inquiries and might refuse to operate if they receive an unfavorable response.

GrapheneOS will display a notification every time this API is utilized, making it simple to identify which apps are accessing it. While the majority of apps do not use it, a few certainly do. I noticed a few apps accessing this API, yet none of them failed to function; they were satisfied with secure booting. However, some apps are more selective, and there is a short list of apps that won't operate on GrapheneOS. It's crucial to test any essential apps before switching to an alternative build like GrapheneOS as part of due diligence. There's always the risk that a future app update might cause a previously working app to stop functioning; this is a definite risk when using any alternative Android build.

Security features

GrapheneOS offers various security and privacy features in addition to its extensive system hardening. Many of these features ensure that the device operates as if it is truly owned by its user. For instance, the provisioning data that comes with Android, which guides the device on collaborating with carriers globally, permits carriers to dictate that certain features, like tethering, should not be accessible. GrapheneOS never managed to implement that part of the system. Instead, there is an option to stop the phone from reverting to older, less-secure cellular protocols.

The standard Android system allows control over certain app permissions, but it doesn't allow users to block an app's network access. However, GrapheneOS offers this control, although network access is initially enabled to ensure compatibility. When network access is turned off, the app perceives the environment as if access is available, but the device simply never connects to a signal. Therefore, apps shouldn't decline to operate merely due to the lack of network access, although they might not function properly.

A "sensors" permission bit regulates access to sensors not covered by other permissions, such as the accelerometer, compass, thermometer, or similar devices. This permission is also activated by default but can be disabled by the owner.

The storage scopes feature allows apps to operate in a sandbox, giving them the impression of full access to the device's shared storage, but they can only access files they have personally created. Similarly, the contact scopes feature permits apps to think they have complete access to the owner's contacts, while actually keeping most or all of that information concealed from them.

GrapheneOS includes fingerprint unlocking similar to standard Android, but with a key distinction: after five failed attempts, the fingerprint option is deactivated for 30 minutes. This allows a device owner to rapidly disable the feature by using a finger that isn't recognized if they are compelled to unlock the device. For individuals with heightened privacy concerns, a duress PIN can be set up; inputting this PIN will prompt the device to instantly erase all its data. It's important to note that this self-destruct option should be used cautiously.

A unique application can evaluate the status of a GrapheneOS device and, by utilizing hardware security features, confirm that the device has neither been altered nor downgraded to a previous software version.

The project regularly releases updates, and GrapheneOS systems installed on devices update promptly. The project upgraded to the Android 16 release in early July, just under a month after Google launched that version. By default, the device will automatically reboot after 18 hours of inactivity to ensure all data is stored (encrypted) at rest; this also ensures the device runs the latest software version.

See also this page comparing a long list of security features across several Android-based builds.

Governance and community

One possible drawback is the unclear development community backing GrapheneOS. While a foundation is established to support this system, details on its operations are sparse, apart from a lengthy list of donation methods. Public records reveal three directors: Micay, Khalykbek Yelshibekov, and Dmytro Mukhomor, yet there is no information available on director selection or fund allocation by the foundation.

The project has numerous repositories with its source code, but there's minimal guidance on contributing or insights into the development community's activities. Some details are available on the build-instructions page. The project manages chat rooms and a forum, though discussions are mainly user-focused rather than centered on development. Contributions to the forum by the project are made through a general "grapheneos" account.

In response to a private inquiry, the project stated that it has ten active, paid developers, with most working full-time. However, it seems that Micay still plays a key role in leading GrapheneOS; at the very least, the project's aggressive stance on the fediverse closely mirrors his previous interaction styles. The outcome if he were to leave the project is uncertain. This presents a potential risk that is difficult to measure.

Configuring the device with GrapheneOS took a few days, primarily focused on replicating the apps and settings from the older device. Time was needed to properly adjust privacy settings and assign necessary permissions to apps. Ultimately, the device functions just as effectively as its previous version, offering all essential features while excluding many unnecessary ones. I am wholeheartedly dedicated to using it and have no plans to revert.

The system is undeniably more secure, even if the unseen hardening modifications have no real effect. The sandboxing is stricter, there's greater control over what apps can do, and there isn't any AI trying to break free.

Naturally, the ongoing issue is that GrapheneOS by itself won't suffice for many individuals, necessitating the introduction of proprietary software. Although the documentation states that Play Store login isn't mandatory, it demanded a login from me, reconnecting the link to Google that GrapheneOS installation had severed. The keyboard doesn't allow for "swipe" typing, so users wanting that feature will probably install GBoard, which comes with its own privacy concerns. The GrapheneOS messaging app functions, but Google's app can filter out some spam, so it might also be worth adding. Some sensible, privacy-friendly weather apps are available on F-Droid nowadays, however, the proprietary ones that compromise privacy have superior access to weather alerts (at least in regions with operational weather agencies) and red-flag warnings. Android Auto is very practical and functions well on GrapheneOS, though it necessitates its own set of special access permissions.

Additionally, there are numerous banking, ride-sharing, airline apps, and similar services that appear essential in today's world. However, each of these app breaches the privacy barrier that GrapheneOS has meticulously built. It's possible to survive and even prosper without them, and we know some who do, yet these tools are popular and available for valid reasons. For many people, it's just not feasible to manage without using proprietary software, much of which is known to monitor our activities and behave in unfriendly ways.

Installing GrapheneOS on a phone ensures awareness of each vulnerability created and encourages minimizing these vulnerabilities as much as possible. When potentially harmful software must be permitted on a device that holds sensitive information, the system will strive to keep that software confined to its designated limits, preventing it from acting beyond its permissions. Installing GrapheneOS aligns a device more closely with the owner's interests, which in itself is worth the investment.

Share

Subscribe

]]>kazani@newsletter.paragraph.com (Kazani)androidprivacysecurity<![CDATA[Which Android Browser is the best?]]>https://paragraph.com/@kazani/which-android-browser-is-the-best YHdfqeYXx3UEt5oOthN4Tue, 22 Jul 2025 13:33:59 GMT

Regular person who follows the crowd? Chrome

Regular person who uses Microsoft? Edge

Regular person who wants no Google or Microsoft? DuckDuckGo Browser

Want the privacy but also want to save trees? Ecosia

Want even more privacy and don't care about trees? Brave

Want maximum privacy? Tor

Too slow? Privacy Browser

Want customization? Vivaldi

Even more customization? Soul

Something like Soul but more basic? Via

Too basic? Opera Mini

Are you a gamer who likes to see news on games? OperaGX

Not a gamer? Opera

Still not enough features? Quetta

Changed your mind and don't want a Chinese browser? Aloha

Want something that won't get any more new features? Arc Search

Want something Firefox-based? Firefox

Not private enough? Firefox Focus

Don't want FF but want FF-Based? Fennec (F-Droid or GitHub only)

Want something else? IceRaven (GitHub only)

Want something from the Play Store? Waterfox

Not secure or private enough? IronFox (GitHub for F-Droid link)

Want something more obscure? SmartCookieWeb Preview (GitHub)

Same obscurity but Chromium? Fulguris (GitHub)

Less obscure but weirder? FOSS Browser

Less obscure and more stable? Kiwi Browser (the GitHub one, not the one on the Play Store)

Less obscure and got a Samsung device? Samsung Internet

Got a Samsung device but don't want Samsung apps? Opera Touch (Galaxy App Store only)

Want something as close to Chrome as possible but as secure and private as possible without resorting to Tor? Cromite (GitHub)

There are many others like Phoenix which has a built-in file explorer and Tempest which has only 1 Search engine (Tempest). If you're curious about any of these, they can be found in the Google Spreadsheet https://docs.google.com/spreadsheets/d/12v8TE3pr74bZR_ExFKjH7oOloOPn68f35TPskNTJbio/edit?gid=0#gid=0 or found through the search engine of your choice.

Always read the Privacy Policy first. You can get a good idea how the company operates by the ways they protect themselves

The best ones I've come across are Cromite, Brave, Soul, IceRaven, IronFox and Via. They all work great for me, everyone is different.

Share

Subscribe

]]>kazani@newsletter.paragraph.com (Kazani)androidbrowserprivacy<![CDATA[13 Open Source Alternatives to ChatGPT]]>https://paragraph.com/@kazani/13-open-source-alternatives-to-chatgpt irZmfQ0qm47xwtiGd7lxMon, 22 Jan 2024 14:05:47 GMT

Good Morning! 😄

Delivering interesting content every single week on Bitcoin, Web3, Security, Crypto, Privacy & AI.

It's FREE, Takes less than 5-minutes to read, and you are guaranteed to learn something.

Subscribe to get valuable News, Useful Resources and Insights every week to your Inbox!

Subscribe

ChatGPT , developed by OpenAI, is a robust generative AI tool allowing users to input text prompts in a conversational manner, receiving detailed responses. While beneficial, it's proprietary and not open source. This article highlights 13 open-source alternatives, each catering to various needs.

Not all ChatGPT alternatives function the same way. Some solutions are purely meant for developers to create their own chatbot on top of it. And, a few others offer a chatbot or demo for you to test.

\ \ https://www.adilkazani.com ChatGPT Guide: Go from Zero to Hero \ \ Most people use ChatGPT but they don't really know what it can do. Here in this guide you will learn: - ChatGPT basics - To start -more............](https://www.adilkazani.com/2023/05/chatgpt-guide-go-from-zero-to-hero.html)

Reasons to Look for Open-Source ChatGPT Alternatives

Relying on any single service is bad for the consumers. The same goes for ChatGPT.

In addition to that, here's why we should look for open ChatGPT alternatives:

I do not intend to say that ChatGPT is bad, or you should stop using it. However, "we" as the users should benefit more from the open alternatives in the long-run.

Not all demos allow commercial use. You need to be careful when using the content from the chatbot demos for some options.

1. OpenChatKit

Developed by Together,

OpenChatKit is a comprehensive alternative to ChatGPT, using the RedPajama model. Visit its GitHub page for technical details.

Collect this post for FREE on Zora Network (only 100 mints available)

Connect Wallet

Collect

2. ChatRWKV

An open-source alternative powered by an RNN (Recurrent Neural Network) language model, ChatRWKV offers a demo on Huggingface. Explore its GitHub page for details.

Developers and businesses can build their chatbots utilizing ChatRWKV.

3. ColossalChat

Part of the Colossal AI initiative, ColossalChat allows you to clone AI models and build ChatGPT-like platforms. While the demo isn't functional, the source code is available on GitHub.

4. KoboldAI

\ \ https://github.com GitHub - KoboldAI/KoboldAI-Client \ \ Contribute to KoboldAI/KoboldAI-Client development by creating an account on GitHub.](https://github.com/KoboldAI/KoboldAI-Client?ref=adilkazani.com)

Geared towards assisting writing, KoboldAI is a browser-based front-end AI designed for novels. It supports various modes and can be explored on its GitHub page.

5. GPT4ALL

\ \ https://github.com GitHub - nomic-ai/gpt4all: gpt4all: open-source LLM chatbots that you can run anywhere \ \ gpt4all: open-source LLM chatbots that you can run anywhere - GitHub - nomic-ai/gpt4all: gpt4all: open-source LLM chatbots that you can run anywhere](https://github.com/nomic-ai/gpt4all?ref=adilkazani.com)

GPT4all is an open-source project enabling chatbots to run locally on CPUs and almost every GPU. Install its desktop application and explore details on its GitHub page.

6. HuggingChat

Leveraging Huggingface's platform, HuggingChat is an open-source ChatGPT alternative. Take it for a spin and explore the source code for customization.

7. Koala

\ \ https://bair.berkeley.edu Koala: A Dialogue Model for Academic Research \ \ The BAIR Blog](https://bair.berkeley.edu/blog/2023/04/03/koala/?ref=adilkazani.com)

Koala by EasyLM is a locally run chatbot built on the LLaMA dataset. While the demo isn't available, check its documentation for running it locally.

8. Vicuna

Trained on top of LLaMA, Vicuna claims to rival ChatGPT in quality. Explore this open-source chatbot on its official blog post.

9. Alpaca-LoRA

\ \ https://github.com GitHub - tloen/alpaca-lora: Instruct-tune LLaMA on consumer hardware \ \ Instruct-tune LLaMA on consumer hardware. Contribute to tloen/alpaca-lora development by creating an account on GitHub.](https://github.com/tloen/alpaca-lora?ref=adilkazani.com)

Alpaca-Lora provides an instruct model using low-rank adaptation, with the ability to run on a Raspberry Pi. Details are available on its GitHub page.

10. Dolly

Dolly is a language model trained on Databricks for commercial use. Find the source code on GitHub and explore the model on Huggingface.

11. H2oGPT

Tailored for queries and document summarization, H2oGPT offers a live demo and source code on GitHub for exploration.

12. Cerebras-GPT

Cerebras-GPT offers open-source GPT-like models with a focus on parameters for improved accuracy and compute efficiency. Find model details on Hugging Face.

13. OpenAssistant

Although the demo is no longer functional, OpenAssistant aimed to provide access to a ChatGPT-like chatbot. Utilize the available progress and source code for further development.

In conclusion, open-source alternatives to ChatGPT offer flexibility, transparency, and customization. Users and developers can choose based on their specific needs, ensuring adherence to language model policies. Explore, modify, and utilize these alternatives to enhance your chatbot experience.

If you're enjoying today's newsletter, why not share it with your friends? They might find it just as informative and entertaining as you do.

Sharing is caring, and by spreading the word about this newsletter, you're helping to support ME and ensure that more great content gets produced in the future. Plus, you'll get to have even more conversations with your friends about the interesting topics covered in each edition.

There are three ways to show me that you enjoyed reading this article:

  1. Share this post with your friends

Share

  1. Subscribe to my newsletter

Subscribe

  1. Collect this post for FREE on Zora Network (only 100 mints available)

Connect Wallet

Collect

I hope this was helpful!

Thank you for reading!

Let’s bust some more in next article. 😄

Related ChatGPT Links:

\ \ https://paragraph.xyz How to Use ChatGPT: A Step-by-Step Guide for Beginners to Enhance Your Conversations and Writing \ \ Unlock the potential of ChatGPT, the AI tool everyone's talking about! Learn how to elevate your conversations and written content with this versatile tool. ChatGPT is the AI tool that's taking the world by storm! With its advanced natural language processing capabilities, it can help you with every](/content/@kazani/how-to-use-chatgpt-for-beginners/index.html)

\ \ https://paragraph.xyz ChatGPT's Chain of Thought Prompting: A Powerful Tool for Cognitive Flexibility and Adaptive Thinking \ \ ChatGPT's Chain of Thought Prompting is a powerful technique that can be used to promote cognitive flexibility and adaptive thinking. By prompting users to generate a chain of related thoughts in response to a particular question or topic, ChatGPT can help individuals to explore complex ideas and ga](/content/@kazani/chain-of-thought-prompting/index.html)

\ \ https://paragraph.xyz 🌟 100+ CHATGPT AI PROMPTS 🌟 \ \ Elevate your digital marketing game with this diverse assortment of AI-generated prompts designed to cater to all aspects of content creation. From email marketing templates to YouTube video ideas, and from ChatGPT SEO prompts to promotional social media posts, I have gathered an extensive array of](/content/@kazani/100gpt-prompts/index.html)

\ \ https://paragraph.xyz 🔍 7 Realistic Ways to Make Money with ChatGPT \ \ The digital age has brought a paradigm shift in the way we create, consume, and monetize content. Artificial intelligence, and ChatGPT in particular, has emerged as an invaluable tool to empower content creators, marketers, and entrepreneurs to unlock new revenue streams. In this article, you'll dis](/content/@kazani/7-realistic-ways-to-make-money-with-chatgpt/index.html)

]]>kazani@newsletter.paragraph.com (Kazani)chatgptopenaiaiopensourcealternatives<![CDATA[How do I make a Bitcoin Private Key offline?]]>https://paragraph.com/@kazani/how-do-i-make-a-bitcoin-private-keys-offline gJUWtTkrcApMf2ZznxGzTue, 26 Dec 2023 12:27:25 GMT

**Good Morning! **

Delivering interesting content every single week on Web3, Security, Crypto, NFTs, Privacy & AI.

It's FREE, Takes less than 5-minutes to read, and you are guaranteed to learn something.

Subscribe to get valuable News, Useful Resources and Insights every week to your Inbox!

Subscribe

In light of the recent controversy surrounding Senator Pocahontas and her proposal to ban the ownership of certain words, I thought I'd create a simple guide on generating a bitcoin private key offline. This way, you can also become a threat to our democracy.

Requirements:

1. Paper

2. Pen or pencil

3. Dice or Coin

4. Calculator

5. Printed out copy of the BIP-0039 protocols word list from GitHub

https://github.com/bitcoin/bips/blob/master/bip-0039/english.txt

6. An analog computer, or hardware wallet.

Collect this post for FREE on Zora Network (only 100 mints available)

Connect Wallet

Collect

Step 1:

A. For a 12 word seed phrase, number your paper 1 through 12, skip a line & number it one through 12 again.

B. Write out “ 1024, 512, 256, etc…” at the top columns of the paper as shown in the image below:

Step 2:

A. Roll dice ( even numbers = 0, odd numbers = 1) or flip your coin ( heads = 0, tails = 1)

B. Record your roll on the paper. Begin at the “1” beneath “1024” & continue to roll and record under each column from left to right.

Step 3:

Once you have completed your rolls on the first row take every column where you recorded a “1” & add the number above it, then add the “+1” at the end of the row & write the number.

In the image below 1 was recorded in 512, 128, 64 & 16.

512 + 128 + 64 + 16 + 1 = 721

Step 4:

Continue this process until you get to the final 4 spaces on the paper.

Note: The 12th word to your seed is your “ checksum” word. This word must be a mathematically compatible fit with your previous 11 words ensuring your seed can function with the BIP-0039 protocol.

Step 5:

A. Before we can find out the final checksum word we must first translate our binary digits to their corresponding BIP-0039 seed words.

B. Open up the word list & write down the word associated with each number that you wrote down.

Ex. (721 = foam)

Step 6:

Continue step 5 all the way to word 11.

Step 7:

A. To find the final checksum word we must first import our 11 words onto our hardware wallet.

(For this example, I am using a COLDCARD wallet MK4 by Coinkite)

B. Select “ Import Existing” on device.

Step 8:

A. Select “ 12 words” on device.

B. Enter all your words up to word 11.

Step 9:

A. To find your final checksum word, add up your 1s and 0s in your 12th row.

B. Get out your BIP-0039 word list & start at word “ 801” (the sum of your 7 rolls) for this example.

Step 10:

Your checksum could be any word between 801 & 816 so we will have to guess each one for the 12th word by inputting it on the device until the device confirms that we found a compatible checksum.

Step 11:

When you find your 12th word your device will confirm it’s a fit by applying the private key into a PubKey private key dataset that you can use to sign TX’s with the device, & broadcast to the internet without revealing sensitive data (like private key) through PSBT’s.

The odds of someone else being able to guess your specific private key is 2048^12

You have better odds of picking out a specific atom on earth, than anyone has to crack that key.

Adding a passphrase moves this from the realm of mathematical improbability & into impossibility.

Generating a key completely offline is something all of us must know & teach for maximum security. We can’t trust generating this on any device, or application, etc which all represent potential attack vectors.

Everyone has a right to preserve the fruits of their labor securely in a monetary standard that doesn’t redistribute your wealth.

The state & banking parasites hate Bitcoin because it’s money that they can’t weaponize against you unlike the dollar.

credit: Why Bitcoin Only

Related Useful Links:

\ \ https://armantheparman.com Make a Bitcoin Seed Phrase from Scratch (Using Dice) \ \ Updated version Making your own private key is a great feeling. I'm going to show you how, and keep it easy to follow. Most people will not go to the extreme of doing everything in this guide, but reading it will give some appreciation to how bitcoin storage works, the level of safety that is...](https://armantheparman.com/dicev1/)

\ \ https://privacypros.io What is a Mnemonic Phrase and BIP39? (2022 Update) \ \ This is a complete guide to the BIP39 and seed phrases. Find out how recovery phrases work in this in-depth post.](https://privacypros.io/wallets/mnemonic-phrase)

https://www.bitplates.com How secure is your passphrase? | BitPLATES® \ \ The BIP39 passphrase (or '25th-word') will keep your Bitcoin safe and secure, for as long as it can hold-off a brute-force attack.

https://enteropositivo.github.io BIP39Colors - BIP39 mnemonic to colors tool \ \ Enter a valid 12 or 24 words BIP39 compatible mnemonic or generate one using an external tool like Iancoleman mnemonic tool BIP39Colors BIP39Colors as Palette Enter 8/16 BIP39colors and you will get back your BIP39 mnemonic. Error: The seed must be 12 or 18 words length BIP39 compatible BIP39Colors offers a BIP39 compatibile secure approach to storing cryptocurrency seeds.

https://3rditeration.github.io Seed Savior: Mnemonic Phrase Recovery Tool \ \ Seed Savior: Mnemonic Phrase Recovery Tool

\ \ https://vault12.com Generating a Seed Phrase using a Calculator. \ \ Understand how to use an offline calculator as mnemonic generator. In this article, we walk though the methodology and tools used in generating a 12-word seed phrase, also known as a mnemonic phrase or sentence, to copy into your chosen cryptocurrency wallet.](https://vault12.com/securemycrypto/cryptocurrency-security-how-to/calculator-seed-phrase-generator/)

\ \ https://blog.keyst.one How to verify the recovery phrase created by dice rolling \ \ When setting up your bitcoin wallet, the most crucial part is the process of creating your recovery phrase. Have you ever wondered where these recovery phrases come from? It's from a random number...](https://blog.keyst.one/how-to-verify-the-recovery-phrase-created-by-dice-rolling-af01c16b765e)

\ \ https://unchained.com Ultimate guide to storing your bitcoin seed phrase backups - Unchained \ \ In this article, we cover how and where to store your seed phrase backups in different bitcoin custody contexts, like singlesig, multisig, and more.](https://unchained.com/blog/how-to-store-bitcoin-seed-phrase-backups/)

If you're enjoying today's newsletter, why not share it with your friends? They might find it just as informative and entertaining as you do.

There are three ways to show me that you enjoyed reading this article:

  1. Share this post with your friends

Share

  1. Subscribe to my newsletter

Subscribe

  1. Collect this post for FREE on Zora Network (only 100 mints available)

Connect Wallet

Collect

I hope this was helpful!

Thank you for reading!

**Let’s bust some more in next article. **

My Previous Articles:

\ \ https://paragraph.xyz IS PRIVACY IMPORTANT TO YOU? 🔒 INTRODUCING "RAILGUN" - Private & Anonymous DeFi \ \ Did you know that there already exists a chain-native private DeFi system on Ethereum, Binance Smart Chain (BSC), and Polygon? The RAILGUN project has developed this private DeFi system, which is user-friendly and operates independently of Layer 2 solutions or bridges, ensuring no compromises on sec](/content/@kazani/introducing-railgun-private-and-anonymous-defi/index.html)

\ \ https://paragraph.xyz Bitcoin Halving: Unveiling the Secret Behind Digital Gold \ \ Discover the intricacies of Bitcoin Halving and its profound impact on the cryptocurrency landscape. Explore the mechanism behind reducing mined Bitcoins by 50%, its timing, and implications for supply dynamics. Learn about the scarcity concept in Bitcoin, its relationship to price trends, and the m](/content/@kazani/bitcoin-halving/index.html)

\ \ https://paragraph.xyz I Have 42 Custom Filter Lists for uBlock Origin Extension You Must Import Right Now! \ \ Are you ready to take control of your online browsing experience? Discover the game-changing influence of incorporating 42 custom filter lists into your uBlock Origin extension. Say goodbye to intrusive ads, safeguard your privacy, and elevate your browsing to new heights!](/content/@kazani/42-custom-filter-lists-for-ublock-origin-extension/index.html)

\ \ https://paragraph.xyz 💎💧YOUR 100X FINDING TOOL - "ALPHANOMICS" (BLOCKCHAIN INSIGHTS THAT MATTER) 💧💎 \ \ 🚀 Want to be ahead of the curve in discovering promising tokens? It's simple: 1️⃣ Get a top-notch tool 2️⃣ Develop a killer strategy That's all you need! Let's explore a set of strategies to apply using one of the top blockchain intelligence tools! 🧰](/content/@kazani/alphanomics-your-100x-finding-tool/index.html)

\ \ https://paragraph.xyz Enforce Privacy & Security Best-Practices on macOS Using Terminal in 15 minutes \ \ Discover how to fortify your macOS privacy and security effortlessly by harnessing the power of Terminal commands. Safeguard your digital realm with these expert tips.](/content/@kazani/enforce-privacy-and-security-best-practices-on-macos-using-terminal/index.html)

...and more!

]]>kazani@newsletter.paragraph.com (Kazani)bitcoinseedphraseprivate keysprivacyhuman rights<![CDATA[IS PRIVACY IMPORTANT TO YOU? 🔒 INTRODUCING "RAILGUN" - Private & Anonymous DeFi]]>https://paragraph.com/@kazani/introducing-railgun-private-and-anonymous-defi jhzpgrHAf9IW0f9BE5cFMon, 04 Dec 2023 09:33:03 GMT

Good Morning! 😄

Delivering interesting content every single week on Web3, Security, Crypto, NFTs, Privacy & AI.

It's FREE, Takes less than 5-minutes to read, and you are guaranteed to learn something.

Subscribe to get valuable News, Useful Resources and Insights every week to your Inbox!

Subscribe

RAILGUN: Enhancing Privacy in DeFi

Did you know that there already exists a chain-native private DeFi system on Ethereum, Binance Smart Chain (BSC), and Polygon? The RAILGUN project has developed this private DeFi system, which is user-friendly and operates independently of Layer 2 solutions or bridges, ensuring no compromises on security.

RAILGUN serves as a privacy middleware available on Ethereum, Polygon, Arbitrum, and BNB Chain, allowing direct integration with applications and wallets. The primary interface for RAILGUN is Railway, an independent wallet enabling private transactions and anonymous swaps through 0x, a decentralized exchange (DEX) aggregator.

Collect this post for FREE on Zora Network (only 100 mints available)

Connect Wallet

Collect

Uniqueness of RAILGUN

Unlike many competitors in the privacy space, RAILGUN operates on base layers like Ethereum and seamlessly integrates with external smart contracts, setting it apart. The project began gaining traction in July 2021, offering users native and hassle-free privacy solutions without relying on external privacy chains or basic mixers.

Project History

RAILGUN emerged in July 2021 with open-source contributions and subsequently introduced the RAIL token via airdrop, establishing the RAILGUN DAO for governance. Notably, it launched the RAILGUN SDK on Ethereum and continued expanding to BNB Chain and Polygon.

How RAILGUN Functions

Shielding Assets

Users engage the RAILGUN Privacy System by "shielding" their assets. This non-custodial process involves sending assets (currently ERC-20 tokens or NFTs) to a public RAILGUN 0x address, which can then be shielded to a private 0zk address via RAILGUN's privacy pool, ensuring token anonymity.

Transaction Mechanics

RAILGUN transactions occur between 0zk addresses or within DeFi, involving encrypted data transmitted to Relayers via the private Waku network, ultimately interacting with the blockchain. Transactions use a UTXO model, employing ZK proofs for anonymity, while revealing only essential details like Relayer and destination addresses.

For RAILGUN-to-RAILGUN (0zk) transactions, the sender, recipient, token type, and transacted amount remain totally private. The publicly visible details are the Relayer address and the destination address, i.e., the respective RAILGUN smart contract (for the given chain).

For transactions involving other smart contracts (e.g., a DEX), the tokens/amount exchanged is also visible. Finally, users can unshield to any 0x address as they can with the typical coin mixer. In doing so, they can anonymously send tokens – linking the transaction to the RAILGUN smart contract address rather than their own.

Integration and Additional Features

Developers use the RAILGUN SDK, employing Adapt Modules to bridge external contracts and the RAILGUN Privacy System. Noteworthy features include gasless transactions, multi-sends, and voluntary disclosure options for creating verifiable transaction histories.

Governance and Tokens

RAILGUN operates with three governance tokens corresponding to specific deployments and DAOs on Ethereum, Polygon, and BNB Chain. Active Governors, staking RAIL tokens, influence governance and earn a share of DAO revenue.

How to Use RAILGUN's Private DeFi System

1. Create a Wallet

Visit RAILGUN's partner project, Railway's app, at https://app.railway.xyz to create a wallet.

https://app.railway.xyz Railway: Private DeFi Wallet \ \ Railway is a private DeFi wallet powered by the RAILGUN network. Shield, Send, and Swap across Ethereum, BNB Smart Chain, Polygon, and more.

2. Two Addresses in Your Wallet

Your wallet will consist of two addresses:

Funds and activity in the 0zk address are hidden from public view such as on etherscan. We call funds in your 0zk address your private balance.

3. Understanding Private Balance

Funds and activities in the 0zk address remain hidden from public view, such as on Etherscan. These funds are referred to as your private balance.

4. Transferring Funds

Transfer funds into your RAILGUN 0x address from any source (e.g., MetaMask or exchange wallet) across supported chains. Simply copy the 0x address generated in the first step and send funds to it as you would to any other address.

5. Import Existing EVM Wallet

You can import an existing EVM wallet by entering your seed phrase. Note: MetaMask will input only the first address if multiple addresses are linked to it.

6. Privately Moving Funds

To move funds into your private balance, initiate the process by pressing the ' Shield' button. Choose the token and amount to shield. Your funds never leave the chosen chain.

Input your password and click 'Shield'. The system generates a zk-SNARK proof, ensuring private ownership of your shielded assets.

7. Additional Features

For enthusiasts, any ERC20 token can be shielded by clicking the '+' icon above your balance and selecting ' Custom Token'.

8. Private Transactions

Once funds are in your private balance, you can privately send them to other 0zk addresses or unshield them back to any 0x public address. All fees are paid in stables/crypto of your choice, eliminating the need for ETH or $RAIL.

RAILGUN SDK: Extending the Capabilities

With the RAILGUN SDK, you gain the ability to interact with any smart contract using your private balance. This means seamless swapping, trading, and earning on your favorite chain's dApps, ensuring complete privacy without compromising user experience or liquidity.

Advantages of RAILGUN Over Other Privacy Projects

Benefits for Users:

RAILGUN for Developers

Seamless Accessibility and Security

Railway DEX, the first cross-contract integration of the RAILGUN SDK, enables token swapping within your private balance with low slippage, utilizing efficient order routing via the 0xProject API.

The availability of Railway as a standalone desktop app, iOS, and Android ensures access to private DeFi from anywhere.

Interested in learning more or joining the world of private DeFi? Visit our Telegram group here!

Conclusion

RAILGUN's success hinges on code evaluation, network effects, and reducing adoption barriers due to fees. Integrating with more platforms and mitigating reliance on Railway wallet could de-risk the project and accelerate adoption, crucial as privacy becomes increasingly vital in the crypto landscape. While facing complexities and challenges, RAILGUN presents a crucial solution for ensuring user and developer privacy in the evolving crypto space, poised for growth with its innovative tools and technology.

The imminent public release of the RAILGUN SDK allows developers to integrate RAILGUN, providing a secure, private DeFi experience without compromising on liquidity or security. Join the world of private DeFi and experience seamless, secure transactions today!