What always strikes me about these lists is that if I had these ideas (and I often do), I'd discard them out of hand as non-viable. Yet people behind these startups don't mind staking years of their life on what is essentially a lottery with rather poor odds, and find passion in doing things which, at best, would make me yawn. Maybe I'm just old, I don't know.
Most startup ideas are non-obvious in their very early phase. If there was already proof one way or another that they were good or bad, they would be already done, or already dropped, respectively.
Also, the founders have much more knowledge about their startup. You have what you got from that one paragraph description you read. That's a huge asymmetry of information.
PG's essay explicitly calls us out:
"At YC we're excited when we meet startups working on things that we could imagine know-it-alls on forums dismissing as toys. To us that's positive evidence an idea is good."
I’ve always felt that for many it’s a hedged situation, the ~75k+ salary, networking, and resume bullet points, represents similar or greater compensation than what they might make alternatively.
A C-level position at a YC startup will grant you a 100k+ salary at a YC company that did manage to make it to a series A afterwards even if your company didn’t make it that far.
Even folks working on the problem sometimes think it is non-viable. I question the thing I'm working on, every single day. That's just how it is, even after you commit.
If one can deliver real value to end-users, and the growth takes a phenomenal turn, then different kind of non-viable thoughts would emerge, I'm sure.
I kinda agree. I just don't see the value in 90% of those ideas even if they are well executed. Maybe I'm getting old too but it seems most of these propositions are based on the premise of making something attractive and getting funding from VCs wanting to diversify their investments and play the lottery.
The lowest hanging fruit in tech has already been taken years ago and the human needs haven't changed that much. We need to communicate, we need to listen to music, we need to move around, etc.
I think most look at the ideas and don't see a safe/profitable path. Most are right but with the right person they can take any idea and make it work. Not because of the big idea but by winning the small battles and the non-obvious battles.
Facebook had a similiar idea to others. But they limited it to Harvard. So anyone who joined was part of a social club that they saw in real life. That gave them trust and people more freely posted. Next they rolled out to only ivy league schools. Next all colleges. Then highschools. When they released to the public the colleges still had semi-private networks allowing them more freedom. Then facebook did something no one thought was possible, they allowed you to login to hotmail import your contacts. Now you are friends with people you may not have talk to in years.
The original idea only shaped a broadvision. Each decision really provided the vision and a path for growth.
The original vision is important. If you tried to take a dating app and follow the Harvard facebook roll-out it wouldn't work because people wouldn't want to be seen as looking for a mate. Dating sites try to connect strangers.
I have feeling, that team is most important here, not starting idea. There is chance, that the teams with currently non-viable ideas will pivot and be successful in the future.
- Cool seeing machine learning being used inside applications (The audio analyzer to separate and boost voices for conference calls comes to mind)
- Surprised at how many "meh" reactions I had. Too many grocery/fashion and not enough bio/energy game changers and I left feeling mostly uninspired by overall visions.
Was talking to some VC folks the other day and asked this same question. The answer I got was under performance, not enough unicorns per investment and the number of companies even hitting profitability was much lower than other sections of their portfolio.
I am currently participating in another accelerator in London called Antler; and I have been exposed to ideas from 70 entrepreneurs. We are presenting to investors this week.
Also I am in YC Startup School too, which allows you to see what's coming up in the world, and that's awesome too.
It's amazing to see that a significant percentage (~30%) of ideas they have come up with have been also thought of by others e.g. from the YC list independently to an uncanny extent.
Honestly it should not have caught me by surprise but it did.
> It's amazing to see that a significant percentage (~30%) of ideas they have come up with have been also thought of by others e.g. from the YC list independently to an uncanny extent.
Imho the concept of multiple discovery [0] applies just as much to innovation as it does to scientific discoveries.
When I advice folks... I always like to say, for any good idea you have, 100 other people have already thought of it. For any great idea you have had, 1000 people have thought of it.
The core difference lies in the execution. If you can execute on even a marginally “okay” idea, you can build on that, but even the greatest idea in the world has no chance without that execution.
Corollary: don’t worry about protecting your idea, worry about how you can be the best person to execute it.
In looking at all the bio-based companies they all seem ... mature. Like, going on their pages and seeing the people there, well, there are not a lot of 'fresh' faces. Most of them are MDs/PhDs or have some years on them. I think thats a great thing!
"is raising a $7.5 million seed financing at a valuation north of $30 million"
"We’re told several top venture capital firms were vying for a stake in Tandem. One firm even gifted the founders a tandem bike, sources tell TechCrunch, resorting to amusing measures to sway the Tandem team. But it was a16z — which has an established interest in the growing future of work sector, evidenced by its recent investment in the popular email app Superhuman — that ultimately won the coveted lead investor spot."
This sounds like what Kevin Kwok was describing[0] in his recent essay, The Arc of Collaboration. Even down to using the Discord metaphor. Kwok says that Slack ended up being more of an exception handler when normal processes break down. Contrast to Discord, which is more of a meta-layer around games and integrates within game platforms. I wonder if Tandem is what will make collaborative communication ubiquitous, i.e. sales talks to engineering talks to design all from within their various environments.
From TC: Tandem provides a virtual office for remote teams, complete with video-chatting and messaging capabilities, as well as integrations with top enterprise tools, including Notion, GitHub and Trello.
MyPetrolPump is genius. It's last-mile delivery for gas (or Uber for gas), but solving a security and logistics problem for a high-value market. Only qualified truck drivers can move trucks, and you pay them per hour. So now you refuel the trucks behind closed doors while the truck is parked overnight and you only have to secure access for the petrol driver.
I can say that Tandem is the one which strikes me the most. I had a privilege to test the product in early beta and you can see the benefit of such solution. Basically Discord for work which I believe will be adopted soon by many companies.
> TrustedFor: LinkedIn is just such an awful platform that there’s space for a startup to disrupt it by just remaking it. TrustedFor is building “LinkedIn 2.0,” a platform for professional profiles that is centered around recommendations from people that the users have actually worked alongside. The startup is leveraging the YC network pretty heavily to get associated companies on board.
Good luck. I'm happy to see someone trying to get into LinkedIn's business. I've used LinkedIn close to when it started, but it is really getting on my nerves.
If they succeed, that could be really interesting on how to compete with an established company where the network effect is their real edge.
The Custom Movement sounded really interesting for a moment. I was already having fantasies about getting a pair of sneakers with thin soles, wide toe boxes, plenty of air holes and one the left shoe one size larger than the right.
Alas. It's just about printing designs on sneakers.
I'm quite curious about Waves. There have been a number of unsuccessful attempts (that I'm aware of) to make a dating site/app with kink matching, but it's hard to keep the emphasis on sexual compatibility from completely overwhelming the other aspects of dating.
Looking at their site, there doesn't appear to be any obvious secret sauce (the app is not available where I am) except perhaps that the range of 'kinks' they're catering to is limited to the mildest and most popular, so maybe they're targeting a slightly more mainstream userbase?
Working in the area makes this text on Tensil’s page me extremely curious: “This gives you the performance of custom silicon at the cost of commodity hardware. With Tensil, you can roll out your dream chip in weeks instead of years, at prices measured in thousands instead of millions.” So they designed some fuse programmable ASIC for machine learning? Setting the fuses makes this ASIC “custom silicon”? ASIC for thousands? Sounds to good to be true.
Tensil founder here. We're a little different from the other ASIC companies because we specialize our chips to individual model architectures on demand. Our value prop is that our chips are smaller and cheaper than high end GPUs while offering the same or better performance on the model you care about.
This is compelling to a lot of hardware companies, where they often have 1 model that represents >50% of their compute workload. Making a Tensil ASIC for that model becomes much cheaper than GPUs at scales of 1000 chips or more.
I'm happy to talk in more detail if this interests you! [email protected]
ASIC set-up costs are only a problem on low volume production runs, and in that case generic purpose hardware is probably a cheaper solution anyway.
If you have some high computation or bandwidth requirements that can't be met with of the shelf hardware, then FPGAs are probably still a better solution.
So if you ask me, Tensil is just trying to ride the AI wave.
I'm quite confused my Tensil, is the training of the model moved to their chips, or the final model? If the former, then are the chips locked in to whatever model they were built for, e.g. this chip only trains a multi-layer perceptron with n layers? Or are the chips re-programmable or FPGAs? If the latter, don't most models run quite quickly on CPU after training anyways?
It seems like they're selling ASICs. It would be useful in something like an embedded system where you train a model beforehand, deploy it, and then let it go. On top of speed improvements, they're generally more efficient, so something running on battery power could last longer.
Tensil founder here: we take fully trained models, training doesn't happen on our chips. The model architecture is hardened but the parameters can be reprogrammed.
Vendr.com: A SaaS to keep track of your SaaS subscriptions.
and a 6.5k or 10k PER MONTH for someone to buy your SaaS subscriptions seems super expensive. I'm not too versed in the SaaS market but is this really something that you need a dedicated "SaaS expert" to manage for you?
Signing up with a typical SaaS service these days typically involves 1) a qualification call with an SDR (wastes 30ish minutes), 2) a redundant qualification call with an AE (roughly 30 minutes where they probe you for price sensitivity), 3) a 2-3 day delay while they draw up a really high anchor for your price, 4) two or three iterations of you fighting the price (wastes an hour each time), 5) you sign a contract with a lingering uncertainty you got ripped off.
Repeat for the 20-30 SaaS providers you end up using. It's a huge time suck for a dubious benefit. I would love to pay to make that go away in the future.
Basically, it saves you at least one potential IT headcount. From my time at Amazon, I remember an internal tool to they had to do exactly this. You put software requests and someone from IT buys it for you and sends you the license and installation instructions.
So if your company is large this becomes a real issue I guess. Also by having a third-party IT vendor, you can buy anonymously. This shields you from unreasonable negotiations and future solicitation.
Here are the ideas that I think have the most potential...
talar: Huge potential here. Cars are on their way out, and scheduled grocery delivery can replace it.
spotless materials: Coatings can make a huge difference in the way we interact with materials... in fact, for the most part, that's all we interact with. better coatings make things better.
encellin: I believe that interacting with our bodies on a more finely tuned small scale is the future.
my petrol pump: I can see a lot of convinience happening here. Recurring customers whos lives are made better.
rejuvenation technologies: I want to see life extension succeed. If they have something that works then that's fantastic for us all.
tensil: Baking AI seems like a good compromise for a lot of reasons. known capabilities, and known weaknesses make for predictability.
I hope they can get this to work! The issue is their paper uses technology covered by other players in the space who have a decade head start and significant research budgets. Maybe they'll get bought out, but my guess is delivery and CMC are hurdles they'll run up against and then have a hard time getting follow-on investors. FIH for this idea is going to be really hard. If they can get this going, I'd love to see adding in additional components of the telomerase holoenzyme.
I hope they lean on the exosome delivery idea discussed within, "In highly preferred embodiments, the delivery vehicle is an exosome". That would really differentiate them from the competition. The real test will be biodistribution in non-rodent models. I will be following closely!
Is there a directory which lists how the YC starutps are doing? As in how many from 2016 batch are still around, and how many didn't make it? I am sure there has to be one around.
wow- I didn't even have the patience to read through all of them. How does YC coach so many? At 15 min a startup, they can cursory mentor 32 startups/day or 160/wk if they do nothing else. That doesn't seem like coaching. More like a certificate school.
[+] [-] m0zg|6 years ago|reply
[+] [-] ovi256|6 years ago|reply
Most startup ideas are non-obvious in their very early phase. If there was already proof one way or another that they were good or bad, they would be already done, or already dropped, respectively.
Also, the founders have much more knowledge about their startup. You have what you got from that one paragraph description you read. That's a huge asymmetry of information.
PG's essay explicitly calls us out:
"At YC we're excited when we meet startups working on things that we could imagine know-it-alls on forums dismissing as toys. To us that's positive evidence an idea is good."
[+] [-] ben_jones|6 years ago|reply
A C-level position at a YC startup will grant you a 100k+ salary at a YC company that did manage to make it to a series A afterwards even if your company didn’t make it that far.
It’s a self accommodating industry (IMO).
[+] [-] ignoramous|6 years ago|reply
If one can deliver real value to end-users, and the growth takes a phenomenal turn, then different kind of non-viable thoughts would emerge, I'm sure.
[+] [-] pier25|6 years ago|reply
The lowest hanging fruit in tech has already been taken years ago and the human needs haven't changed that much. We need to communicate, we need to listen to music, we need to move around, etc.
[+] [-] wolco|6 years ago|reply
Facebook had a similiar idea to others. But they limited it to Harvard. So anyone who joined was part of a social club that they saw in real life. That gave them trust and people more freely posted. Next they rolled out to only ivy league schools. Next all colleges. Then highschools. When they released to the public the colleges still had semi-private networks allowing them more freedom. Then facebook did something no one thought was possible, they allowed you to login to hotmail import your contacts. Now you are friends with people you may not have talk to in years.
The original idea only shaped a broadvision. Each decision really provided the vision and a path for growth.
The original vision is important. If you tried to take a dating app and follow the Harvard facebook roll-out it wouldn't work because people wouldn't want to be seen as looking for a mate. Dating sites try to connect strangers.
[+] [-] lnsru|6 years ago|reply
[+] [-] postscapes1|6 years ago|reply
- Where have all the IoT companies gone?
- Cool seeing machine learning being used inside applications (The audio analyzer to separate and boost voices for conference calls comes to mind)
- Surprised at how many "meh" reactions I had. Too many grocery/fashion and not enough bio/energy game changers and I left feeling mostly uninspired by overall visions.
More power to everyone hustling out there though.
[+] [-] Hermitian909|6 years ago|reply
Was talking to some VC folks the other day and asked this same question. The answer I got was under performance, not enough unicorns per investment and the number of companies even hitting profitability was much lower than other sections of their portfolio.
[+] [-] udkl|6 years ago|reply
Previous discussion : https://news.ycombinator.com/item?id=18710682
[+] [-] firtoz|6 years ago|reply
Also I am in YC Startup School too, which allows you to see what's coming up in the world, and that's awesome too.
It's amazing to see that a significant percentage (~30%) of ideas they have come up with have been also thought of by others e.g. from the YC list independently to an uncanny extent.
Honestly it should not have caught me by surprise but it did.
[+] [-] freeflight|6 years ago|reply
Imho the concept of multiple discovery [0] applies just as much to innovation as it does to scientific discoveries.
[0] https://en.wikipedia.org/wiki/Multiple_discovery
[+] [-] rcarrigan87|6 years ago|reply
[+] [-] not_a_moth|6 years ago|reply
Looks like it ranges 60-80k [after their fee] for a 10% stake.
That feels pretty expensive for an accelerator, especially one that's new, with little network effect yet.
[+] [-] aerophilic|6 years ago|reply
The core difference lies in the execution. If you can execute on even a marginally “okay” idea, you can build on that, but even the greatest idea in the world has no chance without that execution.
Corollary: don’t worry about protecting your idea, worry about how you can be the best person to execute it.
[+] [-] atlasunshrugged|6 years ago|reply
[+] [-] IshKebab|6 years ago|reply
[+] [-] Balgair|6 years ago|reply
[+] [-] elmar|6 years ago|reply
https://techcrunch.com/2019/08/20/tandem/
"is raising a $7.5 million seed financing at a valuation north of $30 million"
"We’re told several top venture capital firms were vying for a stake in Tandem. One firm even gifted the founders a tandem bike, sources tell TechCrunch, resorting to amusing measures to sway the Tandem team. But it was a16z — which has an established interest in the growing future of work sector, evidenced by its recent investment in the popular email app Superhuman — that ultimately won the coveted lead investor spot."
[+] [-] amsilprotag|6 years ago|reply
From TC: Tandem provides a virtual office for remote teams, complete with video-chatting and messaging capabilities, as well as integrations with top enterprise tools, including Notion, GitHub and Trello.
[0] https://kwokchain.com/2019/08/16/the-arc-of-collaboration/
[+] [-] whoisjuan|6 years ago|reply
[+] [-] pgt|6 years ago|reply
[+] [-] elkynator|6 years ago|reply
[+] [-] chichikid|6 years ago|reply
[+] [-] cbanek|6 years ago|reply
Good luck. I'm happy to see someone trying to get into LinkedIn's business. I've used LinkedIn close to when it started, but it is really getting on my nerves.
If they succeed, that could be really interesting on how to compete with an established company where the network effect is their real edge.
[+] [-] Tharkun|6 years ago|reply
Alas. It's just about printing designs on sneakers.
[+] [-] mcthrow|6 years ago|reply
Looking at their site, there doesn't appear to be any obvious secret sauce (the app is not available where I am) except perhaps that the range of 'kinks' they're catering to is limited to the mildest and most popular, so maybe they're targeting a slightly more mainstream userbase?
[+] [-] lnsru|6 years ago|reply
[+] [-] 323454|6 years ago|reply
This is compelling to a lot of hardware companies, where they often have 1 model that represents >50% of their compute workload. Making a Tensil ASIC for that model becomes much cheaper than GPUs at scales of 1000 chips or more.
I'm happy to talk in more detail if this interests you! [email protected]
[+] [-] LeonM|6 years ago|reply
ASIC set-up costs are only a problem on low volume production runs, and in that case generic purpose hardware is probably a cheaper solution anyway.
If you have some high computation or bandwidth requirements that can't be met with of the shelf hardware, then FPGAs are probably still a better solution.
So if you ask me, Tensil is just trying to ride the AI wave.
[+] [-] ben_e|6 years ago|reply
[+] [-] christophclarke|6 years ago|reply
[+] [-] 323454|6 years ago|reply
[+] [-] CabSauce|6 years ago|reply
[+] [-] fybe|6 years ago|reply
and a 6.5k or 10k PER MONTH for someone to buy your SaaS subscriptions seems super expensive. I'm not too versed in the SaaS market but is this really something that you need a dedicated "SaaS expert" to manage for you?
[+] [-] evtothedev|6 years ago|reply
Signing up with a typical SaaS service these days typically involves 1) a qualification call with an SDR (wastes 30ish minutes), 2) a redundant qualification call with an AE (roughly 30 minutes where they probe you for price sensitivity), 3) a 2-3 day delay while they draw up a really high anchor for your price, 4) two or three iterations of you fighting the price (wastes an hour each time), 5) you sign a contract with a lingering uncertainty you got ripped off.
Repeat for the 20-30 SaaS providers you end up using. It's a huge time suck for a dubious benefit. I would love to pay to make that go away in the future.
[+] [-] gitgud|6 years ago|reply
Wouldn't you just need them to arrange and liaise with the vendors at the beginning? Once the plans are finalised, why is Vendr.com still needed?...
[+] [-] dang|6 years ago|reply
[+] [-] whoisjuan|6 years ago|reply
So if your company is large this becomes a real issue I guess. Also by having a third-party IT vendor, you can buy anonymously. This shields you from unreasonable negotiations and future solicitation.
[+] [-] morcutt|6 years ago|reply
http://www.siliconhillsnews.com/2018/05/02/meta-saas-acquire...
[+] [-] pontifier|6 years ago|reply
talar: Huge potential here. Cars are on their way out, and scheduled grocery delivery can replace it.
spotless materials: Coatings can make a huge difference in the way we interact with materials... in fact, for the most part, that's all we interact with. better coatings make things better.
encellin: I believe that interacting with our bodies on a more finely tuned small scale is the future.
my petrol pump: I can see a lot of convinience happening here. Recurring customers whos lives are made better.
rejuvenation technologies: I want to see life extension succeed. If they have something that works then that's fantastic for us all.
tensil: Baking AI seems like a good compromise for a lot of reasons. known capabilities, and known weaknesses make for predictability.
[+] [-] DrAwdeOccarim|6 years ago|reply
I hope they can get this to work! The issue is their paper uses technology covered by other players in the space who have a decade head start and significant research budgets. Maybe they'll get bought out, but my guess is delivery and CMC are hurdles they'll run up against and then have a hard time getting follow-on investors. FIH for this idea is going to be really hard. If they can get this going, I'd love to see adding in additional components of the telomerase holoenzyme.
Edit - Their IP position is actually pretty strong and they are already thinking about delivering the other holoenzyme components: https://patents.google.com/patent/US20140242155A1
I hope they lean on the exosome delivery idea discussed within, "In highly preferred embodiments, the delivery vehicle is an exosome". That would really differentiate them from the competition. The real test will be biodistribution in non-rodent models. I will be following closely!
[+] [-] pgt|6 years ago|reply
[+] [-] glebedel|6 years ago|reply
"hire CTO" as a goal
Team capabilities star ratings...
[+] [-] sidcool|6 years ago|reply
[+] [-] jerrre|6 years ago|reply
[+] [-] Finbarr|6 years ago|reply
[+] [-] lucasverra|6 years ago|reply
[+] [-] vadym909|6 years ago|reply
[+] [-] adrianN|6 years ago|reply