ptrenko

Yes and it ends eventually. A single AI makes it out of here

I know so many people who have Google/FB jobs straight out of college. It depends on your perspective, but its not rare.

I assume anyone who wants to startup is atleast a top 5% engineer. Otherwise its just a bad punt. You have to be terribly good to be in the startup biz

So AutoML is not smart yet? Its just about throwing more compute?

Yeah I've learned a lot about boostrapping working at a startup in college. The physics is completely different.

You have to be very very efficient. You understand why Heroku costs 3x of AWS. You understand that raising money is not necessary for every project. For most tourist entrepreneurs its just cargo culting.

But now I think I wanna spend a good 10 years in corporate before even thinking of doing anything else. Make that moolah first.

Why do people care about free speech so much?

When things get really bad, smart people simply pack their bags and leave.

Politics is unpredictable and there's no time to waste arguing for free speech.

People have no idea why free speech is useful.

They just want free speech.

Makes sense that this would evolve naturally. At that depth only the groups that communicated this way could coordinate and survive

Sorry not $10, more like $7

Source: https://www.bloombergquint.com/coronavirus-outbreak/not-rs-5...

India announced a Rs 1.7 lakh crore economic relief package >ensuring cash transfers of Rs 500 a month, and food >security. However, the package is less than one percent of >the country’s GDP.

If this is the new reality of the world we live in, what do you think will be the new protocols for international travel? I imagine a standardized heath test without which you can't board a flight?

I think I'll faint seeing AI progress anymore.

I didn't even think this was possible. Have people ever done this manually before? Like without AI?

I'll tell you another reason why people underestimate difficulty of startups.

The probabilities are completely uncertain. You just never know how big it can get or how likely it is.

If you are optimistic like I was, you "hope for the best" since that one guy did it right? In essence, you don't know how hard the problem is.

In the corporate world you do know the probabilities, so you get scared by it.

TL;DR: Once you know the true probabilities you'll know which is a better option

Read this analysis: https://www.myroadtofire.com/blog/how-engineers-can-fire-in-...

You have to be an L5 engineer at Google. And I know a guy earning $700k per year at 32-33 in ARM. Plus other VPs. Depends on your vantage point, but starting a startup is way harder. Personal experience here.

I think thats an efficient way of testing if a country is actually improving its situations and getting out of the lockdown. Even traffic data could be useful.

If pollution is still low, the country is not actually back to work yet

They are arguing?

There has never been a time when I've expressed an opinion in my name and not regretted it later. Stay offline or anonymous. More importantly, stay clean

Okay so I'll hang it then. Brain overloading.

People have it really easy there. I'm from India and the needy are getting $10 per month.

I think its $70 per month accounting for PPP, but even then the difference is massive

I wonder what allows Zoom to have a $20B market cap with such silly security loopholes? Like isnt that a consideration in enterprise

Haven't had the time to read the paper but spoke a lot about something similar with a friend.

In terms of crisis, the medical system just doesn't scale. You think leaders were worried about running out of ventialtors?

No! You can start pumping them in hundreds of thousands if you really need it.

The problem is that you have only so many doctors and trained nurses to work with. Limited capacity. Training takes 10 years.

No scalability.

It doesn't matter where the tech is right now. The world will push 100% for whatever can be done via apps, sensors and unskilled workers.

Check out Jonathan Rothbergs vision for a diagnostic kit by everyones toothbrush

Next: How to run wordpress on ENIAC

I just hang it. Does that work?

I think I'll go live in a cave if I hear privacy debated once more!

I'm too inundated with this stuff!

People give all kinds of wrong advice when you start out: Take risks when you're young. Fail fast, break things.

Here's the reality I've experienced as a startup founder. Its not the holy truth, but its been my life and my experiences.

If you shoot an arrow even a few degrees off-center, it hits somewhere far from the board.

Similarly the decisions you take early on are so much more critical to your life.

As an Indian student these are all of the decisions you take:

Age 15: Science, Commerce, Arts

Age 17: Medical, Engineering

Age 18: Computer Science, Electrical, Mechanical, Civil engineering etc

Age 22: US Masters or Local Job

Guess what?

If you want to realistically be a millionaire by 37-38 you have to take 5 correct decisions in a row (Science, Engineering, CS, Masters).

You have to do 5 very difficult things to be a Google level engineer.

Even if you do four out of five correct after 20 years your life will be in a far lower orbit with respect to finances. Its sad but true.

This is why you should do your startup after you clear this early minefield of difficult decisions. Once you can comfortably have a good job at your whim. I'd advise people to startup only after 30.

And remember, one misstep and the outcome can be wildly different after 20 years!

Trying to reduce hallucination in NLG systems. The idea is straightforward enough. Whatever is a parameter of the sentence: tone, nouns, verbs, actions, adjectives need to be the input parameters.

Everything else glues together the inputs to form gramatically correct sentences.

Unrelated question for the AI experts here: What's the potential for AutoML solving problems with just X and Y. I understand that you'd need to search a large subspace but at what stage is it currently and can it solve problems as complex as the one here?

Haven't read it through in detail, but what would be cool is an AI layer that suggests an image automatically. A bit far fetched I agree. Good work!

I remember seeing a graph (from Paul I think):

graph1: linear + exponential s curve

When you zoom out the first graph, the exponential part starts looking linear

graph2: linear + slightly less linear + exponential

Oh wait, it is an exponential curve right? D'oh

Strange. I'd expect less traffic from novelty type e-com sites ...

Just a question: Has anyone figured out serverless ml predictions yet?

It was too painful. Spent 6 years or so. I;ve given up. Honestly corporate salaries will earn you 3-4 million by the time you are 37 or so. Why bother with rinky dink startups anyways?

Personally loved win 7 the most!!