Hacker News new | ask | show | jobs
On Full-Stack Startups (medium.com)
24 points by peter123 3981 days ago
11 comments

I don't know who this is, but the following isn't about them or their story but the format which I tried to and failed to follow.

1/I'm sorry I couldn't finish it,

2/let alone refresh enough times to get all the content to load without error.

3/Disseminating that much continuous information in that format is atrocious.

4/I'll post a lot of stuff, but it will be so spread out that you won't know

5/If I made any good points or not

6/But the sheer barrage of noise will convince you I know what I am talking about

7/???

8/Profit!

Couldn't finish it either. I thought that maybe there would have been commentary after a few of the tweets, but the format is bad.

I did see a few good points (Fullstack helps you understand legacy practices and why things were the way they had been..), but a lot was definitely lost in all the noise.

Yeah F this, I'm not going to bother reading this if they can't bother composing it in a coherent way. Whatever it is, certainly isn't important enough for me to bother.
I don't know why people do this. I guess it gives off the feeling of it being very important like they are rushed and HAVE TO TELL YOU THIS RIGHT NOW.

But it's not shareable unless you're going to retweet 40+ tweets (are people doing this?). You have to wait for someone else to collect your thoughts for you.

And at the same time it isn't as permanent as actually taking the time to write a blog post so it's easy to retract later.

Trying to load this page crashes Safari on my iPhone 6+ every time. Am I the only one?

It seems like more and more bloated JavaScript is being jammed into every page load these days...

Apparently I have no clue what full stack means. I thought it was being able to manage front-end, back-end, and (maybe) devops code.

After this, I think it's a reference to a new type of Medium post whereby an author stacks full the post with Tweets.

I might be the only one thinking this wasn't a painful but kind of novel/artistic presentation. But onto the substance: no industry needs more of a full stack startup than US hospitals and the healthcare industry. There are some that have started, but piecemeal solutions are excruciatingly difficult to develop and sell effectively.
Got about halfway through before I realized this guy is using Twitter to write what should be a blog post. While I'm sure his message is solid, I'm not sure how much I respect someone's tech advice when they don't seem to understand the technology they're using in the first place.
I enjoy ideas. Here are the ones from the tweet storm:

• Full stack law firm: template all contracts, law APIs as core, hyperdeflate legal costs.

• Full stack clinic: mobile EMR/EHR, qself, genomics, telemed. Employ MDs w/ tech skills. Accept insurance, prefer cash subscriptions.

• Full stack architect: put @vannevartech APIs at core of new construction co. Start with unmanned buildings like datacenters to derisk. Ultimate goal = hit enter to build a building w/ drones + prefab.

• Full stack accounting firm: Given bank account, auto prepare it all: tax to diligence to S-1 w/ legal sign off. (@xero could do this).

• Full stack restaurant: Mobile order/pay/reserve, A/B tested dynamic menus w/ supply chain integration, robot prep.

This is possibly the worst presentation format in a field crowded with worthy contenders. I particularly like the Knight Rider-like green thing that just goes back and forth across the top of the browser.
This format is painful to read, so I'm copying the content here. This is a series of tweets by https://twitter.com/balajis

--

If you are using software to go after a legacy physical vertical, we are most excited if you do it "full stack" [0].

With @counsyl, we found that one cannot live by bioinformatics alone. So built a full clinical lab + software + national salesforce.

Here is a video of our robotic genome factory: Also: insurance coverage, clearances, clinical integration... [1]

The reason to do full stack is that it's hard to replace just one layer of an outdated legacy stack. CAC & integration costs can kill.

For example, consider "restaurant powered by technology" vs "tech for restaurants". With code, reinvent industry and make better margins.

You cannot automate something till you've done it manually many times. You cannot abstract an API till you've seen many specific examples.

Need to build a site of the scale of http://Amazon.com - and encounter all the problems - before you could do http://aws.amazon.com .

Reinventing agriculture? Build new kind of farm, w/ solar panels & self-driving tractors. Control all factors, show that code cuts costs.

More examples of full stack startups.. Education: [2], Autos: [3], Hospitality: [4]

By doing it end-to-end, you see things. First, sometimes a good reason beyond tradition (eg regulatory, logistical) for legacy practices.

Second, your control of adjacent layers may eliminate the initial problem you were trying to solve. Analogous to: [5]

Third, solving whole problem builds respect in your vertical. Now you can recruit best people. Their experience + your code = invincible.

Fourth, your margins improve nonlinearly once you control the full stack. Simply having all orders in DB from day one is transformative.

Remember: for many legacy cos, IT = cost center. They only adopted the code. You were born in it, moulded by it. [6]

"Muni commissioned a $3 million audit..after two years..figured out where people were getting on the bus." [7]

A full stack technology-powered bus service (like @leaptransit, soon to relaunch) can dump that from database for $0 in <1 second.

First full stack entrant into a given vertical is formidably protean. Morph your product just by hitting keys. Ex: [8]

Why not more full stack startups? Perhaps #1: recruiting difficulty. Need experts in many different areas. Must know what you don't know.

A good way to get around this: gain vocabulary. Find best cos in your vertical. Read bios of execs on LinkedIn. Google new terms. Repeat.

Now you know regulations/acronyms/terms like CPT, 510(k), KOL. Repeated mentions increase relevance; prioritize accordingly.

Email the best. Read blogs, get context. Now you know kung fu, or at least know you need to hire kung fu master. [9]

In near term, expect iconic full stack companies in areas like biomedicine (3DP/genomics/qself), finance (bitcoin), aviation (drones).

Uber/Lyft, Airbnb, Tesla just the start. Poetically: almost any vertical regulated by a TLA is ripe for an eventual full stack co IPO.

Under appreciated aspect: full stack company can manage compliance more easily once beyond early stage. Hit enter, dump DB, submit form.

An ideal full stack co has: a) git/DVCS at the core b) domain experts c) domain monetization knowledge d) a burning reason to fix things

As @justinkan wrote recently [10], unit economics dominate physical businesses. So monetization usually starts early.

In re: funding, full stack startups can be more expensive than pure code. Best candidates often 2nd time founder with domain knowledge.

But are thinking hard on alt funding models, for both open source [11] & full stack cos [12].

With that said, b/c VC = power law [13], full stack startups arguably easier to fund b/c easier to envision 1B+.

Finally, the best full stack founders have ambition to think big and the pragmatism to start as small as possible (but no smaller).

1B+ lifetime funding doesn't phase us. 1B+ up front usually a bad sign. Careful staging + thought on monetization = pragmatic ambition.

To sum up: @cdixon & I love full stack cos. If you're a founder w/ company built on git + TLA knowledge + pragmatic ambition, do ping.

By request, some specific examples of full stack startup ideas in a few verticals: law, medicine, architecture, restaurants, accounting.

Full stack law firm: template all contracts, law APIs as core [14], [15], hyperdeflate legal costs.

Full stack clinic: mobile EMR/EHR, qself, genomics, telemed. Employ MDs w/ tech skills. Accept insurance, prefer cash subscriptions.

Full stack architect: put @vannevartech APIs at core of new construction co. Start with unmanned buildings like datacenters to derisk.

Full stack architect: Ultimate goal = hit enter to build a building w/ drones [16] + prefab [17].

Full stack accounting firm: Given bank account, auto prepare it all: tax to diligence to S-1 w/ legal sign off. (@xero could do this).

Full stack restaurant: Mobile order/pay/reserve, A/B tested dynamic menus w/ supply chain integration, robot prep: [18]

If goal is full stack, always talk to execs in field early on. Few words can save years of work, identify key cost centers & hard parts.

Note: can start all of these as "just" a new clinic/restaurant/accountant/architect/law firm. Think big, start small. Prove, then scale.

[0] http://cdixon.org/2014/03/15/full-stack-startups/

[1] http://www.technologyreview.com/video/521121/innovators-unde...

[2] http://techcrunch.com/2014/03/18/former-googlers-altschool-r...

[3] http://www.forbes.com/sites/jimgorzelany/2014/02/27/why-tesl...

[4] http://www.fastcompany.com/3027107/punk-meet-rock-airbnb-bri...

[5] https://en.wikipedia.org/wiki/Telescoping_series

[6] https://www.youtube.com/watch?v=F157geaXp_w

[7] http://www.sfweekly.com/2010-04-14/news/the-muni-death-spira...

[8] http://www.wired.com/2014/02/teslas-air-fix-best-example-yet...

[9] https://www.youtube.com/watch?v=6vMO3XmNXe4&feature=youtu.be

[10] http://justinkan.com/exec-errands-post-mortem

[11] http://startupboy.com/2014/03/09/the-bitcoin-model-for-crowd...

[12] http://jerzygangi.com/why-silicon-valley-funds-instagrams-no... (404)

[13] http://blakemasters.com/post/21869934240/peter-thiels-cs183-...

[14] http://blog.csdisco.com/2013/05/01/the-future-of-legal-techn...

[15] https://en.bitcoin.it/wiki/Contract

[16] https://www.youtube.com/watch?v=W18Z3UnnS_0

[17] https://www.youtube.com/watch?v=Hdpf-MQM9vY

[18] https://www.youtube.com/watch?v=S0GKIQdpscg

On 497 requests for 2.1MB of data (cache disabled)...
Holy fuck....I did not realize it was that bad...
What a terrible medium with which to convey that message. I'm sorry to see this kind of excuse for a story for the second time this week.
What a ridiculous format to publish an article in. If you can't be bothered to format your stream of thoughts into proper sentences and paragraphs, I don't think there's much value in me reading it.