suhail

gpt zero, which detects ai writing, thinks it’s 100% written by ai

I actually asked Marc Andreessen this question a few years ago while making a browser.

Me: “why did you decide to make links blue?”

Marc: “I sure as hell wasn’t going to make them pink.”

Me: “what about green?”

Marc: “ew”

I re-did the course with this 2022 version. Highly recommend it :)

Two factor was shipped a while ago. We'll work on SOC-2 later this year. Seems straightforward. I've done it before.

We plan to relentlessly invest in security.

I understand for some trusting us will not be enough and for others it will take time.

Big announcements soon.

We're delayed by the chip shortage and are expecting 34 huge servers to land around ~mid-July but they've been delayed since April. Right now we've only got 1/3 of our capacity available which isn't a whole lot.

Answer from Marc Andreessen: https://twitter.com/pmarca/status/1481095915427577856?s=20

Congrats! It was a fun many years competing & excited to see the space validated in the public markets.

> "Also, history tells us when companies say they wont be evil... they mean "not yet". It is a middle man power play pure and simple."

That's factually incorrect that, that always happens and I am not sure why you think we're grouped into that without any evidence or a basis of reasoning.

Fwiw, unlike the examples above, we have permission to your data (if you give it to us) but we don't own or control any of it. We don't store it elsewhere. We don't sell it to the highest bidder. We don't mine it for some other purpose.

Let me be clear: we plan to charge a subscription. That's our only business model.

I think we can improve privacy and security for most users who have trouble managing it. For instance, we can patch Chrome zero-days (many have occurred this year) a lot faster for everyone.

What we offer today is a faster way to use applications that do own your data so you can be much more productive and hopefully enable a new set of applications never before possible.

We help improve decentralization of the web over duopolies like Apple/Windows. We make the browser more powerful, not less.

We improve the market share of Linux as a consumer computing OS as it underlies our tech.

In time, we might be open to people owning their own hardware and running Mighty on it but I think a lot of people will prefer we make it "just work" for now. I don't view either world as mutually exclusive.

If there's an opportunity to research making things trustless, we'll work on that.

Linux has low market share, Apple won't let you on non-Apple hardware, and Windows has high fees to license in that way.

The only path is the browser and most desktop apps are web apps.

Feel free to read my post that goes deeper about our thinking: https://blog.mightyapp.com/mightys-secret-plan-to-invent-the...

He might also be talking about many more ideas too: Dropbox, Boom, Lambda School, and another dozen ideas within YC that all seem surprisingly possible. You can criticize him for having skin in the game but you could equally commend him too: he puts his money where his mouth is.

Over 60% of the US population does growing every day.

I am not sure if you read our website but a couple of comments:

1. We use GPUs and many vCPUs actually help with multi-processing since most browser tabs will peg a single CPU

2. A lot fortune 500 companies have adopted "browser isolation" products that do something similar but aren't focused around speed.

We did write something very similar: “Mighty does three things to protect your data: your data lives on our secure servers that are audited by 3rd party security firms, it has tight control in terms of who can access it, access is heavily audited and logged, and your most sensitive data is also encrypted.”

Can we do better? Sure and we will. I am really committed to that as a security conscious person.

Most people want an experience where the underlying OS and the application (the browser) interoperate seamlessly versus having to tame two desktop experiences. The primary application people think is slow is their browser by a wide margin so that's where we decided to focus as more native desktop apps become web apps. That focus lets us constrain the problems we get solve vs boiling the ocean with all of Windows.

Fwiw, we started by streaming Windows and pivoted away.

It's not clear to me that Shadow's business is sustainable. Windows licensing alone for virtualization across end-users if you buy from a reseller is $11/mo/user alone. I only know because we tried and became a reseller briefly. They also seem to use consumer GPUs that violate NVIDIA's licensing and agreements. Maybe they know something we don't.

Fwiw, we had 5 customers pay $30/mo in the last 12 hours who have been trying Mighty for a few weeks.

Believe me, I was skeptical too. I remember sitting in a car driving back up from YC with Michael Siebel asking him: "Hey man, do you think I am absolutely nuts thinking people would pay for a browser that's FREE? That's an idiotic idea right?" and, of course, he encouraged me and I am still feeling pretty encouraged based on talking to users and seeing the revenue/usage/praise 18 mo later.

We have a lot of work to do and I am pretty embarrassed of what we've got still but it felt right to get public about it.

Mighty is: https://www.notion.so/Mighty-is-hiring-945d3168d3e34a37883ca...

Software engineers anywhere in the U.S. or Canada.

Mighty makes Chrome faster & use 10x less memory.

Since the media likes to roll this way, I'll post the part Eric buried at the end:

"I talked to Wennmachers for over an hour for this story on the condition that I wouldn’t quote her. She asked me to write that she strenuously disagreed with many of this story’s characterizations and facts."

It won’t. For users, it will simply make it harder to move to anything not owned by Google.

If you can’t move bookmarks, passwords, extensions, history, it requires more effort for the user to switch.

Not too mention, interop to sync on mobile.

Buried in depths of their documentation? It’s also worked for years too. Google has docs that are completely outdated and neglected. Chrome Extension docs didn’t get updated for 5+ years until recently.

Why put it in the console then? Why have tool tips that say “Access to this API is enabled.”

There’s no improvement to security. This will be a cake walk to reverse engineer.

I don’t see how my point is invalid.

Worked for years. Even showed up in the GCP dashboard as having access. Had no idea it “private”. We just plugged in credentials from Google as per chromium build instructions. They knew we and others were using it because there are stats.

Now we are given 2 months to develop a completely new solution.

I wrote one of the first OpenSocial wrappers so that making apps across various social networks was faster. Not all social networks perfectly supported the APIs as spec'd and had quirks:

https://code.google.com/archive/p/opensocial-framework/downl...

I made roughly 3 applications using it so that I could be faster than all the other developers building apps. Lots of people downloaded the framework when it was opened and some even paid for private copies of it early on.

Congrats Josh :)

Why not use Roboto which is shockingly similar?

Congrats, Dan! :)

Exciting improvements I am glad to see: Intel QSV-accelerated VP9 decoding, Support AMD AMF encoder on Linux (via Vulkan)

Does this work on Google Chrome?

Stripe

More interesting: Why did Instagram instantly get a million users upon public release in the AppStore? How did people discover it?

If I were you...

1. Start with something simpler & work on building up your fundamentals. Build momentum & confidence.

2. Recognize your struggle is part of the journey of learning. When you read and don't understand a word, you look it up. Take the time to fight through the abstraction & learn some core concepts when you don't know what something is. It's ok to get distracted & go down a mini rabbit hole.

You may need to learn 5 things that seem unrelated to the task at hand. Those 5 things aren't a waste, you'll likely find the concepts useful later. I always did.

Consider not using some frameworks to remove some of the magic.

Build something you want so you finish it.