HNHacker News
TopNewBestAskShowJobs

richardfeynman

467 karma · joined September 16, 2011

submissionscomments
richardfeynman··on Boston Dynamics’ bipedal robot Atlas can now grab and toss
Would be interesting to combine a Boston Dynamics-type robot with (1) a GPT-like language model, (2) deep learning algorithms that analyze human movement (a la Dall-E but for human motion rather than images), and (3) text-to-speech, so that you can say "dance like a child" and it can generate a child-like dance.
richardfeynman··on Playing Zork with AI-generated imagery [video]
I got access to LaMDA using Google's AI test kitchen. It's pretty underwhelming compared to chatgpt. I imagine google feels what they felt when jobs released the iPhone.
richardfeynman··on Ask HN: Are there any good open source text-to-speech tools?
Wow, that's actually exactly what I was thinking. I will check that out! Thanks so much, I wish I could buy you a beer.
richardfeynman··on Ask HN: Are there any good open source text-to-speech tools?
This is correct. Hopefully whatever improvement children get in their imagination by listening to these stories will offset whatever minor regression in language skills a child may experience due to this semantical humor, which indeed is intended to follow the form of a typical child's mistake.
richardfeynman··on Ask HN: Are there any good open source text-to-speech tools?
I'm defending myself vigorously because I truly wasn't trying to promote myself, despite what you claim. I took time out of my busy day to explain on this forum my use-case for text to speech, which combines two AI technologies and happens to involve kid's stories. I did not ask people to follow or subscribe to the podcast - that is you putting words in my mouth. I thought this audience would find this use-case interesting because this is Hacker News and I hacked together this solution. I also started my answer with "This doesn't answer the question" because it's not a literal list of TTS technologies; instead, it provides an interesting example of how I used one such technology.
richardfeynman··on Ask HN: Are there any good open source text-to-speech tools?
Well, to state the obvious, OP chose not to google it and to ask HN instead, which is why I provided a technology recommendation and an example of a solution I hacked together using the technology.
richardfeynman··on Ask HN: Are there any good open source text-to-speech tools?
This is such a cynical, unhelpful response. Someone asked HN for recommendations for text-to-speech software. I provided a recommendation and explained my use-case. I wasn't trying to grow my audience by posting this comment. I was trying to explain how I combined two new AI technologies to create a new offering. Of course I could be wrong, but based on my decade-plus experience on Hacker News, I think that this is precisely the kind of project PG intended to be discussed here when he created Hacker News.
richardfeynman··on Ask HN: Are there any good open source text-to-speech tools?
While I don't think it's grim to show ads to kids, particularly for relevant products, yes, I can and will rethink the revenue model.
richardfeynman··on Ask HN: Are there any good open source text-to-speech tools?
I'm not exactly trying to grow my kids stories podcast by plugging it to a developer audience. My comment is relevant because it's a cutting edge use case of text-to-speech.
richardfeynman··on Ask HN: Are there any good open source text-to-speech tools?
On the one hand, there's clear empirical evidence from parents that this helps improve kids' imagination and storytelling ability, and on the other hand we have your pure conjecture that "spoken audio which lacks emotional depth" can harm kids. I'm not buying it, but even if it were true it's pretty clear by the rate of improvement in voice cloning that soon there will be more emotional depth in this form of audio.
richardfeynman··on Ask HN: Are there any good open source text-to-speech tools?
That's not true. There's joy in making kids happy. Here's actual feedback from kids:

"Hi richardfeynman and Merry Christmas! Carissa enjoyed her story and decided after that she needs to submit another one She couldn’t tell a difference. My husband and I could tell it was different, but still pretty impressive how the AI works! We will share with friends this week!" "Thank you so much for the story of Anders and his goat Gizmo. We love Maked up Stories and listen to one every evening, so it really made Anders’s day to have his own story. We have shared it with family and friends."

richardfeynman··on Ask HN: Are there any good open source text-to-speech tools?
@wheelsatlarge the maked up stories podcast has 6 million downloads and this is growing quickly
richardfeynman··on Ask HN: Are there any good open source text-to-speech tools?
Are you kidding? there's tons of joy. I have a backlog of hundreds of kids story requests and now instead of being able to satisfy one kid per day I can satisfy as many as I want. moreover, kids can't tell the difference and love the stories.
richardfeynman··on Ask HN: Are there any good open source text-to-speech tools?
This doesn't answer the question, but I thought it might be relevant to mention here that I've been using chatGPT + resemble.ai to create what I believe is the first kid's stories podcast created entirely by AI.

Here's how it works:

- Kid requests a story about a, b, c on www.makedupstories.com

- chatGPT generates the text for a story, summary, and title

- we send this to resemble.ai (sounds like Tortoise TTS would work just as well), which has a clone of my voice

- the audio file then gets sent to anchor.fm

you can listen to example episodes here on Spotify: https://open.spotify.com/show/6liL4T3kJf1scHq134s0mJ

And here on Apple podcasts: https://podcasts.apple.com/us/podcast/kidscast-kids-stories-...

richardfeynman··on The Magic of mRNA Will Push Medical Advances for Everyone
+1 to this comment. If I've learned anything in the past two years, it's that özlem türeci and ugur sahin are prone to understatements rather than overstatements. There is a serious argument to be made that nobody in the 21st century has saved more lives than these two, and it's clear they're just getting started.
richardfeynman··on Show HN: Kids Stories Podcast Using ChatGPT and Voice Cloning
Hi HN, I'm the host of a popular kids stories podcast called Maked Up Stories (>6 million downloads). My biggest problem is that I get way more story requests than I can handle. I just cloned my voice and started inputting story requests into ChatGPT to automatically generate a kids stories podcast. I'd love your thoughts.
richardfeynman··on Ask HN: What is the most impactful thing you've built?
As a pandemic project, I started a kids stories podcast with my then 5 year old son. Two years later, it has over 5.5 million downloads and I get letters from parents telling me how the podcast has impacted their kids lives. I actually haven't posted an episode in over a year, and need to get back to it!
richardfeynman··on Intelligence Beyond the Brain [video]
16 minutes in and I'm hooked. Very interesting!
richardfeynman··on Billionaire Peter Thiel refused consent for sprawling lodge in New Zealand
There is actually a lot of building going on in Wanaka in general, and in neighboring Hawea and Albert Town, so I believe the local council is willing to approve the right structure in an undeveloped area, at least in principle. There is also another lodge called Edgewater that is not too far from Thiel's proposed development site. So it's not impossible that something eventually gets approved after some iterations.

Put aside Thiel's reputation. The difference between what the council is approving and Thiel's lodge, is that the developments being approved have a certain architectural style. The style features lots of open spaces, compartmentalized sections within houses, and large glass doors. The local counsel feels this blends in better with the surrounding area. You can see an example of this kind of Wanaka architecture here: https://www.masonandwales.com/work/beacon-point-house.

Now I don't know if the local council would approve any building at all in Thiel's proposed site, but I do know they physically walked by to see what the proposal would look like from the track by the lake -- so they definitely took the proposal seriously. This may be something that eventually gets approved, but with a different architecture and more muted footprint, and possibly on a slightly different location on the same patch of land.

We are proud of Wanaka for the school system, how great it is for families, the awesome year-round outdoor sports, its low crime-rate, its tranquility, the natural beauty, its wonderful seasons, and the fact that other Kiwis come here to vacation in summer. Businesses here are owned locally.

Nobody here wants to live with underground bunkers next to them. The whole idea of building a bunker here in Wanaka to survive a nuclear blast is silly anyway. It's akin to trying to build a house on a mountain to survive the possibility of a 10,000 foot high tsunami. The odds of a 10,000 foot high tsunami are tiny, and if there were indeed a 10,000 foot high tsunami, the world would have bigger issues than the location of your mountain house. Likewise, if Wanaka, a small town near the center of the less populated island of a friendly, South Pacific island nation, surrounded by over 2,000 kilometers of sea on all sides, ever becomes the epicenter of a nuclear explosion... well, as a species we'd have bigger issues than whether a billionaire's getaway is built 10 feet underground as opposed to at ground level. We all have to accept some level of risk in life.

In my view, if Thiel changed his lodge proposal to look less bunker-y and more like typical Wanaka luxury lodges, the council would approve the structure, even if he built it in his desired location. He should also ditch his Japanese architectural firm and go with a local one. Just my two cents.

richardfeynman··on Billionaire Peter Thiel refused consent for sprawling lodge in New Zealand
Here's a picture from my house right now with a view of where Thiel is proposing development vs. where development is actually happening

https://imgur.com/a/pwxIOEn

richardfeynman··on Billionaire Peter Thiel refused consent for sprawling lodge in New Zealand
I live in Wanaka, from San Francisco. There's one side of the lake where there are lots of houses and hotels, where it would make sense to have a lodge like this. The side he wants to build his lodge on is the other side. People here are very particular about preserving the natural beauty, which is one consideration that went into this decision. Another consideration--which may or may not have affected this decision--is that many people are sore at Thiel for getting citizenship after just spending a few days here. The normal process take at least 8 years.
richardfeynman··on Richard Feynman – Session IV (1966)
There is audio of this. Ralph Leighton has it.
richardfeynman··on Optimizely to be acquired by Episerver
Thanks for making these points. They seem less hyperbolic and more grounded than your earlier claim ("it was one of optimizely's best decisions") and I appreciate that change in tone.

You characterize my arguments as theoretical, but they are empirical. Optimizely's valuation didn't increase after the decision to kill self serve was made, and VWO was able to build a successful business without killing self serve.

Also, as you now agree, optimizely would have had a better source of enterprise deals with self serve intact. That's important.

Stepping back, every SaaS business has issues converting self serve users to enterprise. Fixing those issues takes hard work, which optimizely was not willing to do. Similarly, every business has higher churn in monthly self serve plans than annual enterprise plans. Fixing this takes hard work too. Things were not unsolvable, optimizely simply had no appetite for solving them.

To address your other points: 1. I donbt that labor costs would have made a difference, but say they would have. Was anything stopping optimizely from reducing costs to have overseas development in india? 3. Your comments about the various foci at different rounds of funding is irrelevant. I'm not arguing that building an enterprise business is bad, I'm arguing that killing self serve was bad.

Ultimately, as Dan predicted in 2013, Optimizely died of indigestion rather than starvation; the market was always there, and still is.

Hope this helps clarify my position, even if we still disagree. If you send me an email, I'd be happy to help squash your skepticism about the $100k too.

PS. I don't think that killing self serve was the only thing that led to optimizely's demise, but I do think it was a factor.

richardfeynman··on Optimizely to be acquired by Episerver
First off, love the username :-).

You argue that pivoting from self-serve to enterprise was one of the company's best decisions. If that were true, Optimizely would be considerably more valuable now than it was four years ago, which is manifestly not the case.

Moreover, if the self serve model was indeed fatally flawed, then it would not be possible for anyone to build a profitable business serving this segment. However, a competitor, VWO, showed that it is possible to build a large, profitable business using this model.

Finally, let's say my prior two arguments are wrong, and the unit economics of the self serve business were fundamentally flawed--an assertion I challenge--the self serve business would still have been useful as a source for attracting future enterprise customers. When Optimizely cut the self serve plans, it also killed its largest source of enterprise deals.

richardfeynman··on Optimizely to be acquired by Episerver
I worked at Optimizely from before its series A in 2012 until the end of 2016, so I have a unique perspective on this. For most of the time when I worked at Optimizely, the company was all the rage. It appeared at the top of most "hot startup" lists, the Glassdoor reviews were 5/5, revenue was skyrocketing, and for a period in 2014 it became the fourth most valuable YCombinator company (after Stripe, AirBnB, and Dropbox). Of course, Optimizely's success was't guaranteed. In 2015, the company abandoned the self-serve market that had driven its original momentum and pivoted instead to vague and indefinite enterprise offerings that were (and are) hidden behind schizophrenic marketing, an impossible sales process, terrible customer service and a general approach of trying to extract the maximum amount of money from clients rather than providing them with value. From 2015 on, everything (including the internal culture) became mumbo-jumbo, a cloud of dishonesty. I used to be able to explain what Optimizely did to my grandmother; now I don't even really understand it myself.

Th Episerver acquisition is indeed a bad exit, and I think I will lose >$100k in stock I exercised (which is OK, I'll be fine). But I hope all readers will take from this saga a lesson in humility and the pitfalls of intellectual dishonesty and hubris. Just because your startup is skyrocketing isn't enough. Success is not guaranteed. Your company's leadership needs to be honest with itself, which Optimizely's leadership was not. They need to be humble and work hard, which Optimizely did not do.

richardfeynman··on Mind the Secret Bunker
I've found a bunch of secretish bunkers and tunnels in San Francisco. They're all over the place starting at ocean beach and going towards GG Bridge.
richardfeynman··on Instacart paying 80 cents an hour because worker received a large tip
I'm also pretty sure that they say that 100% of the tip goes to the driver when you place an order.
richardfeynman··on InVision has no physical headquarters and all 700 employees work remotely
Being remote helps InVision build better products because it forces us to use our own tools.

You mentioned that it's much easier for people to grab a colleague in person and show them something on a whiteboard versus having a virtual meeting - and use this to conclude that it's easier to build products in person versus remotely. Perhaps this is true, and perhaps it isn't true; but as a virtual company InVision is forced to find product solutions to these problems, which leads to innovation. That's how being a remote company helps us build better products.

To continue with the whiteboarding example: We created a tool called Freehand that allows you to have virtual whiteboarding sessions that are more flexible than physical whiteboards. Multiple people can whiteboard at the same time, which is harder in person. You can add comments to the whiteboard, create shapes, use colors, and more. Freehands are saved, so you can look back on them later more easily than a physical whiteboard. I think it's more powerful than an actual physical whiteboard and have seen non-remote companies adapt it. You can try it out yourself at: https://freehand.invisionapp.com/

Obviously, not everything about our product is perfect today, but I hope this explains how being remote leads to product innovation.

Disclosure: Obviously, I work at InVision.

richardfeynman··on UPS’s $20B Problem: Operations Stuck in the 20th Century
I wonder if Amazon will buy them.
richardfeynman··on A Linguistic Enclave in Central Europe
If it's any help, I speak pretty good dutch, some german, and some czech (which is similar to polish) and can only understand about 1/4 of this. Any native dutch speakers can feel free to chime in, but I don't think speaking dutch is a major advantage in understanding this.
← PreviousPage 8 of 9Next →