One of the most important lessons I learned as a founder is that everyone will give you advice, and the advice will often conflict with other advice, and that's OK.
The trick is to look at the source of the advice and think about why that advice worked for them. What aspects of their situation are similar to yours?
So all advice is good, provided you use it all as inputs to your decision making and consider the context when you decide which advice to lean into.
A few years ago before planning our wedding, my mother said to me _You have to listen to advice, but you don't have to take it._ I don't have the best reputation for always listening - her point was that it's polite to listen and then it's up to you to judge whether the advice is right for you and the situation at hand.
Well said! People often ask me for my founder perspective, and I always say the following at the end of such meetings - This is just my point of view biased by my personal experiences. You should talk to more people and hear their points of view. And in the end you should make your own decision.
Exactly because of the same reason you write about.
Someone giving advice usually does not have all needed context of your situation. At the same time they are not going to give you all the context they were in while taking similar decision.
Even if you would deal with the same people as that person, those people might just not like you and they liked him/her. Where someone giving advice will think they are just like that for everyone.
Even advice about "great restaurant" is probably bad, because well they were there a while ago, maybe chef will have a bad day or you will get a rude waiter because you will sit in a different spot, maybe they were there with great friends and you want to go with a girlfriend. It might turn out that restaurant was not that great when you arrived.
Well yes "Sturgeon's law - ninety percent of everything is crap.". Online reviews are useless.
I get to know about things like series from friends or sites. But I don't have any expectations and quickly judge if it is crap or not, well with "it is a crap" as default and expect to be positively surprised.
There are some useful nuggets here, but every company and founder are different. Having run three different accelerator programs I have learned to STOP assuming what makes a good or bad idea. Some of the companies I thought were certain for failure have raised multiple rounds, found product market fit, and are growing just fine. The first 18 months of a startup should include one thing, "GET SOMEONE TO PAY FOR YOUR PRODUCT/SERVICE!" Everything else is just iteration.
Good execution on a weak idea can rapidly be redirected to a better idea. If you can't execute, the ideas don't really matter.
I guess it depends a bit on what you mean by "unpromising". One person might see the product idea and think "those guys are gone is 6 mo", while another might see the team coming together well and think "that team is going to do something interesting"
Also, in my experience most people are bad at evaluating these things early on.
It's a good question. From what I've seen a lot of it has to do with their network and the founder's ability to sell the narrative that what they're building is something no one else can build and/or execute. In those early days of a startup investors look at the team more closely than the product. If you're able to show that you're focused on revenue from day one it tends to shift the leverage. Very few investors are funding ideas anymore.
Promising/unpromising is a judgement call here. If you thought something was a great idea but you were wrong, there's a potential failure mode for a promising startup. And the reverse would give you a successful one that you found unpromising.
That's what the poster you're replying to is saying - it's hard to be consistently correct on if an idea is good or bad.
Depends on what you mean by inferior. If you mean inferior as in quality, then your link doesn't support you:
> It refers to the argument that software quality does not necessarily increase with functionality: that there is a point where less functionality ("worse") is a preferable option ("better") in terms of practicality and usability.
It's a rephrasing of KISS. This isn't about getting something quicker out the door.
Yeah...I'm skeptical to be honest. If you're running Chrome on a server, you have to fetch the website to the server, then send back to the client. Now I suppose the server could be in some datacenter with really fast internet and include some aggressive caching, but that's still two trips instead of one.
I suppose if you're running major web apps in the browser, this could be a good idea. With the rise of WebAssembly, you could end up with something like a video editor or photoshop in the browser. And yeah, if big apps in the browser becomes a thing, people will want to run the big app on slower computers. But it'd have to be one beefy server to make the latency of streaming worth the performance speed up. Maybe if you keep common sites hot in the browser JIT and share it across users it'll be really fast but holy security issues Batman.
I'm not super pessimistic about the idea. If there's one thing I'd bet on it's internet speeds increasing while compute speeds plateauing. But I'm not sure I'd bet my own money or 5-10 years of my life on the idea.
It surely will pivot. Low-latency high-fidelity streaming doesn't come handy, and a lot of hard engineering work. But if worked, a lot of applications. Like you said, 100Gbps NIC between your data and video editor is much better than 1T fast local storage editing rig with 10Gbps NIC to the archive. Citrix worked really well in industries that need tighter control over your digital environment. Oculus Link just started to support WiFi, and would be cool if it can go over WAN.
There are some competitions, such as Parsec or Stadia, but if you can have an higher-frequency use case to iterate on, why not go with that?
Besides complicated apps, I think this could be useful for dumb web pages that are making a bunch of requests. Instead of having all those requests go from the browser to the provider to the pile of services, advertisers, and spies, and back again, they'll just go from the Mighty data center to the provider and back. I'm thinking FB and pretty much any news/old media site.
Why would people want this? With the recent concerns of censorship and getting locked out of platforms this just feels like handing even more power to some central authority that can suddenly decide who can and cannot browse the web.
For most people that’s an implementation detail (that they are unaware of), combined with the overall lack of IT hygiene: I don’t see this being an issue (even if your points are very valid)
With web apps becoming more complex by the day, having the ability do all the heavy work on a remote server and stream the resulting “webpage” back to a weak receiver with little processing power will unlock new ways of doing things: one can imagine a very slim screen that, as mentioned in the thread already, is running the full Adobe Suite from this remote server.
Great battery life and responsiveness without needing an expensive GPU / CPU on the device itself.
Maybe testing? Companies like Browserstack offer that sort of thing as their core business.
Say that you just made a big frontend change to your website. Before you release it, you can open it on the last 10 major versions of desktop and mobile Chrome/Firefox/Safari/etc. Any rendering failures that you find are the sort of bugs that can be very frustrating to debug from occasional user contacts.
But it can be tricky to automate the detection of visual bugs. And what if you have to make a few clicks and execute some javascript to get to the screen that you want to test?
Showerthoughts: why not stream the entire OS at this point, with a dedicated storage? Like an EC2 with a desktop environment that you can access in a VNC-like way but with the resolution, FPS and input responsiveness similar to what Stadia/Mighty offers.
Which will allow you to play Cyberpunk (Stadia) + browsing the web (Mighty) + do whatever else you want, like using After Effects (?).
Since you'll need an OS (a minimal one but still) on the client side, you can reduce the bandwidth required by sending paint commands to draw the UI locally (on the client) and just send the resources on the wire as well as updates. [0]
I've been using Shadow for a year and that's exactly what it does. It's a standard windows box and you can install whatever you want on it, and there's no usage limits. I use it to play windows games on my mac without fussing with dual booting. I played through the Witcher 3 entirely on that machine. :)
Non action games are playable over Wifi, but on a wired connection it's much better. It does have its own client and disconnects if you are idle so I don't think you could run it as a server, but I never tried.
The only odd issue I ran into is that it supports a virtual USB device for pass-through which is cool, but you can't use a fancy multi-button mouse as that USB device because the mouse is "special" and it's always a 2 button virtual mouse but the xbox controller worked just fine.
They do have a fairly slow rollout of new hardware updates and new country support at this point, and seemed to be in danger of going out of business for a while, but so far it's still there. You can always hook up something like Dropbox to back up work in progress.
You use it for gaming? What was the input lag like for the Witcher 3? I imagine you wouldn't be able to play certain genres of games (like online FPS games for example) due to high latency.
That's what Citrix were doing in the enterprise space back in the 1990s. And yes, it felt like magic back then, but it was expensive as hell. I think that's the reason it didn't become "mega" successful. The technology itself was great. I'd demo it to management, they'd get how it made my job easier and allow people to work remotely, but the $$ investment in licenses and servers would immediately shut the conversation down. They'd use it in certain places, like if they had a very old / very new / very hard to configure app, they could just stream it into regular Windows NT machines. They weren't brave enough to shift desktop budget to thin clients and repurpose the aging desktop fleet into dumb Linux terminals. (And actually, the whole Microsoft Client Licence made that almost as expensive as running Windows clients).
Some companies dipped their toes in the water, but I never found a single company that went all-in.
In 2010ish I worked briefly in a school in the UK (ages 11-19) which had 3 or 4 racks of servers running Citrix XenApp, serving Office etc. and general educational and administrative applications to ancient laptops and desktops which had been (very inexpertly) reconfigured as thin clients.
It ran like crap (probably didn't help that re-provisioning a server was defined as hot-swapping disks around and running NewSID or whatever it was called) and of course in large parts of the school the WiFi just wasn't quite good enough. I got a lot of calls from PE teachers trying to take the register from the far corner of the playing field.
It was an extreme case of CV-driven development, I was brought in to backfill when the IT manager who had introduced all this walked out one day. He stood up at about lunchtime, grabbed his coat and said "I'm going now". The team thought he meant for lunch...
By the time I was leaving under a year later, they had brought in another IT manager who (rightly!) said "what's all this nonsense" and rolled out proper clients imaged with MDT.
AWS Workspaces got big during the pandemic. I know atleast one mega consultancy corp which rolled out like 20k workspace instances when their employees moved to WFH.
First thing I thought of was Sun Rays and the magic of taking your smartcard around with you and wherever you say down your desktop would follow you around. Magic!
Ah yes. Sun. A wonderful company. It was like knowing a guy who could write poetry in 15 languages, perform magic tricks, breakdance, recite Pi to a thousand digits, solve a rubix cube in 3s and yet couldn't find a job.
I worked with those guys on a large project. They were amazing. I loved their Sun Rays. But if you spoke to other people in the IT community they generally never heard about their tech. "Uh. Don't they sell servers that don't run Windows?"
I'm working with a number of clients that are doing this. Either an outsourced vmware horizon deployment or Windows Virtual Desktop. Local hardware is only thin clients. No servers in-house, no desktops, no laptops. It isn't as good as local hardware, but it's good enough for most business applications and even video conferencing.
I liked the thread a lot, but as with all things startup I feel a lot like this is trusting a very successful horse race gambler with his perfect method to track the form.
I have a feeling with these things timing and luck is everything no matter how well you follow what worked before. I bet some startups do almost the opposite of this advice and still end up doing well just because the idea and timing were so good.
/21 dont listen to these advice because that only worked one time for one person (myself)
I'm wondering if startup gurus know their advice won't help anyone but still give them because that's what people want to read. Or if they don't know this won't help anyone
My optimistic view is that I think it's more that they are providing a data point rather than a conclusion. If every founder writes their experience, then maybe readers will have a large enough sample to find some signal from the noise (luck of individual founders).
I think this is also why its so important for founders to talk about their general background, professional experiences, and product domain. Understanding the constraints the advisor is working with helps the advisee effectively process or ignore their advice.
> Rolling out of bed in my pajamas and getting back to work with things precisely as I left them was the most underestimated superpower that brought me joy, focus, and speed—I had forgotten how much goes into getting ready for work.
There will be so much resistance to going back to the office.
It really depends on the person. The people who naturally focus and can ignore distractions excel at WFH.
A second category of worker benefits from the rituals of getting ready in the morning, context switching to office mode, seeing their peers in person (however briefly), and sitting down at a separate workstation that is dedicated to work and nothing else.
In my experience managing mixed on-site and remote teams, most people assume they’re in the first category but many people eventually discover they’re actually in the second category.
WFH is great for some, but actually quite difficult for many others.
> The people who naturally focus and can ignore distractions excel at WFH.
Depends on whether there are more distractions at work or at home. Some live alone and have no distractions at home, but at work must try to ignore some really loud people in an open-office and/or engage in constant, impromptu, relatively unimportant meetings at work.
People that can focus instantly on demand and ignore distractions excel at this type of office.
Also, people assume everyone in the world have a bare minimum workspace at home to work distraction free - a basement, or home office or at least a small nook. But there are so many wide variety of people in the world who don’t have such luxuries, live with extended families and had to WFH. I doubt they were productive as such.
Linux TTYs is how I do it. If you press Ctrl+Alt+2, you can open a new terminal and run a new graphical interface. If you use a tiling Window Manager, you can have different "desktops" for every project. By switching you TTY, you move to that work environment where you have left your editor, tabs open, database editor, servers, etc...
I’m in the process of leaving a company who has sent half a dozen surveys to employees about how they think return to office should work, but not a single survey about what employees would like to see from a long-term WFH strategy.
It seems some hope to bait and switch their employees back into office and then say WFH isn’t allowed anymore.
It's gonna be really hard though for many companies who have invested a lot of time, effort and prestige into new office buildings to show off their growth, if those buildinga are gonna stay mostly empty. I can't help but think that dynamic is going to play out in many places.
I also think there are some obvious benefits to physical office space, so I hope I'll be able to combine the two. Like every other week or something would work nicely.
Suhail seems to be spending a lot of time trying to establish himself as a thought leader. How does that square or not square with the goal of giving one's business the best chance of success?
The two themes I'm noticing from this list are "Lean heavily into a founders/startup community" and "Ask tough questions often". There are several bullets around coaching, bouncing ideas off of other (external) founders, leaning into internal cofounders strengths/passions, etc. I would generalize to say these things are not specific at all to the first 18 months at a startup (or startups in general). Leaders should be communicating with other leaders, teammates should be relying heavily on other teammates, all humans should seek counseling (via therapists, coaches and friends). A significant part of this process is knowing when to just smile and nod through unsolicited feedback/shitty advice. This is also totally unrelated to startups.
To his second theme (ask tough questions often), it's just the common trap tons of startup founders seem to fall into. "We don't know if users will pay for this" should be one of the first questions you answer, way way before "How will we scale" or "What tech should we use". Managing worry is identical to asking tough questions (why are we worried about this and how do we overcome the worry). Mitigating risk is identical to asking tough questions (what could cause us to fail that is in our control). A part of being a leader is being unafraid to ask questions that need answering, even if their answers may be painful or scary. Most people aren't capable of looking hard at all of their possible failure scenarios honestly, and even less are able to objectively move forward with the findings of that examination. It's a tough ego game to play.
mixpanel: "Powerful, self-serve product analytics to help you convert, engage, and retain more users."
I'd love to read about founders that actually make something. In days of yore, practically any startup actually made a physical 'thing', unless they were busy writing some useful workstation software. Basing an entire economy on internet advertising and related issues strikes me as a risky bet.
To be fair, probably due to offshoring, about the only useful startups I've personally run into for the last decade or so have 100% been specialized medical hardware. The FDA can quickly become the main focus of your life.
I was at a start-up with a website that had exactly one page. The page had a photo of a bank vault, a company logo consisting of 2 letters, and a phone number.
That was perfectly fine. The business was not a website.
Alright – I don't know what that business actual was, but surely it was a business – that's great. That's one type of a business.
Another type of a business is software products. A business like that can basically be a website. Or more fancily, a web app. Or a mobile app, or a desktop app, or a CLI, or a smart TV OS, or anything else written with code that gets used by people.
Now, people need different things, come from different backgrounds, just behave differently. Analytics software simply helps you learn how your thing actually gets used. Do people drop off somewhere in some flow? How do people use this feature you just built? How many users do you even have? Plug in an analytics service's library and find that out with minimal effort. The basic advice for startup founders is: talk to users. That's qualitative research – product analytics is just the quantitative counterpart. Both have value.
The business was software products, built to order, delivered on CD-R media, starting in 2005. The company grew pretty well, doubling in size every 20 months, eventually to be acquired by a large defense contractor. The website probably wasn't needed at all. Getting customers was simple: look for a SBIR contract that is interesting, then bid for it. Analytics software would have angered the customers enough to kill the business, and it may have been prohibited by contract or by law.
I've seen other software and mostly-software businesses where analytics software was out of the question. One example was a company that made an underwater device. Another example was a company that made an OS that got used in military equipment. (add analytics software if you want to live in prison) Not everything is on the internet.
As an early employee of several small startups that have grown to decent sizes, I'll never take the advice of a successful founder (I'll use it when it works).
After hiring hundreds of people the most knowledgeable people were the ones who worked on successful teams at failing companies, and very few founders go through that experience in the same way as non-founders do.
There are good pieces of advice in founders’ self-reported experience, but it needs to be taken as a distorted data point rather than absolute truth.
Founders tend to rewrite history in ways that benefit themselves presently. It’s human nature, but it’s finely honed in founders who are pitching their personal brand on social media. Even the humble, self-deprecating anecdotes are usually carefully filtered in ways that make them look good in the present, or as something they can wear as a badge of honor for overcoming.
I worked for a small startup that nearly died because one of the founders made some really bad decisions. That same founder now writes a lot of Medium thinkpieces about how to be successful, including advice to do some of the things that were clearly not at all beneficial to our startup.
Similar thought from me. My personal feeling is that the best founder advice comes from people who failed at least once (preferably more than once) before succeeding. And the best employee advice comes from people who've worked at a combination of "good" and "bad" companies. It's hard to have enough perspective to formulate good advice when your sample size is 1 (which could have been mostly luck).
Really, I think some twitter accounts are more vapid feel-good nuggets than actual advice (including this one)
I've checked the thread and while there's some materiality to it, it's not anything I haven't heard elsewhere. I'm also wary of founders that might have "gotten lucky" (a weak project in SV is much easier to get financed than a more solid project elsewhere), I'm skeptical of that founder's current project (looks too niche to be sustainable)
I follow Suhail, and he was the founder of a competitor of mine, the stuff he's working on now is really interesting and a lot of the technical challenges the team are working on are fun and interesting.
He's not someone I consider to be full of bullshit, but his thoughts on startups are as valuable as anyone else's with a few extra grains of salt that are worth considering.
I'll second this one as a founder and early employee at a couple of the recent IPOs. The definition of "successful" here must be that the product and revenue succeeded, not that the team did well within the company. Otherwise, you're just filtering for politicians.
The reason this works is that you're identifying the individuals who are able to make success happen while having the deck stacked against them. People who have this skill and are in a successful company tend to have better experiences and can be more valuable. But it's also harder for you to identify.
The advise on how to take advise. In that case I would need to see advise takers that has failed a few times taking advise from successful founders giving advise before taking your advise.
Folks in this thread are assuming the author is trying to present a failproof method for creating a successful startup. But i don't think this is true. It seems more like simply words of advice.
With words of advice, you take what you can when it applies. But words of advice are hardly comprehensive or sufficient.
4/ Think of a way to make your users have some skin in the game enough to yell at you to make your product better. Charge or trade for it early on. Waiting for your product to be "good enough" reduces the amount you'll learn each day. The first set of users paid me $20 on Venmo!
This will vary by type of market you are pursuing. Some things are harder to charge for and take longer to get a first sale, but, yes, you need to be charging. Otherwise, you are "playing house" (a thing they say at YC) or you merely "have the trappings of a business, not an actual business" (that's my typical phrasing for the phenomenon).
In a nutshell: The difference between a business and a hobby is paying customers.
8/ Be married to the problem, not the technology.
There are some great videos from YC that also make this point and I think it needs to be said more often.
There are also some great nuggets about self-management that can be useful to anyone doing anything hard in life or living through a crisis. Some of these resonate with me as a former military wife who raised two special-needs kids mostly alone as the husband was often gone.
13/ If you're worried about something that will cause your inevitable demise, use my patented Threshold of Worry (tm):
1. Set a quantitative value for the worrisome issue.
2. If it's above the value, worry!
3. If it's below the value, focus on the next risk & ignore
I have a longstanding policy of "bread and circus." If you can work on the problem, work on the problem. If you can't work on the problem, feed everyone and keep them entertained so they aren't freaking out, panicking, fighting, etc.
17/ When building a startup:
If you have fear, de-risk by talking to users.
If you have uncertainty, build a prototype to rapidly rebuild your conviction.
If you have sudden doubt, sleep. Try again tomorrow.
When my kids were little and I was chronically short of sleep, I learned that eating something, drinking something and/or taking a short nap could be the difference between feeling like everything is overwhelmingly impossible and feeling like "The sun will come out tomorrow."
19/ I was ruthless about ensuring I had as much deep work time as possible. Live somewhere boring. Eliminate meetings. Don't meet investors if you're not raising. Build a rhythm each day that enables the largest chunk of hours to get great work done. Users will notice your pace.
This is gold. Stop whining about how you have a phone addiction and you can't turn your phone off because all your friends and relatives will be mad if you don't answer it and so on. Stop claiming that all the shiny tech we live with is controlling your life and you are a victim of circumstance. Start taking control over your time and doing what you need to do to carve out time to think, to work without distraction, etc.
We live in an incredible era with amazing things. But you do need to pick and choose and not feel obligated to be plugged into everything all the time. You do have a choice in the matter.
The trick is to look at the source of the advice and think about why that advice worked for them. What aspects of their situation are similar to yours?
So all advice is good, provided you use it all as inputs to your decision making and consider the context when you decide which advice to lean into.