ngould

You're a dark pattern

OP here. Did not mean to imply that web5 sprung out of web3 in any sense. Really I just found web3 to be a useful reference point for explaining SSI. Lots of people understand how web3 works at this point, but there's way less mindshare around the idea of a digital wallet that actually holds credentials and not just private keys.

To that end, I'm generally happy to support the hype, and hope this stuff gets more attention from the web3 lot.

Bingo! =)

It's a little bit tongue and cheek. (Read the post.)

Definitely possible that Apple will win the identity wars. That said, I don't agree with the characterization of web3/web5 as "having to trust nobody". If anything, all the efforts around identity are meant to allow for bringing IRL notions of trust onto the internet. In other words, the whole pseudonymity thing is not a result of using blockchain, but just the fact that nobody's bothered to add identity verification to a lot of stuff happening in web3. That's changing though.

We hope to ultimately build a Slack/Teams alternative designed for rich, asynchronous human interactions that encourage deep work.

All I can say is keep going. I've been feeling the absence of this for a while, and have taken to making occasional videos on Loom to supplement live meetings and Slack messages. But it's a disjointed experience, and Loom is not trying to be a collaboration tool at all. It would be great to have a solution that does the whole thing end-to-end, with a focus on being more async-friendly than Slack.

By the way, I downloaded the PingPong app, and it seems great. Are you selling subscriptions, or just putting the free version out there for now?

P.S. If you haven't read this article on Figma's growth strategy, I'd take a look. There are some parallels with what you're doing -- i.e. SaaS, collaboration-focused, and attempting to dislodge incumbents with a similar list of features, but less focus on collaboration pain points. https://kwokchain.com/2020/06/19/why-figma-wins/

Electric AI | New York City (NYC) | Software Engineer, Engineering Manager (Data + Full-stack), Product Director | Full-time (Onsite) Electric is the world's first all-in-one, real-time IT support solution for small and midsize offices. Through a chat interface, personalized service and flat-rate pricing we keep your email, computers, Wi-Fi and software running smoothly at a fraction of the cost while eliminating headaches normally experienced with traditional managed service providers. Behind the scenes, we're building out a hybrid human/software platform to resolve and execute IT tasks with maximum efficiency and automation.

Electric AI is backed by Bessemer Venture Partners, Primary Ventures, and others. We're currently post- Series B, and hiring across the board for engineering roles. Our stack: React/Redux running microservices on AWS with Python/Serverless. Running Snowflake/Looker/dbt for data analytics & data science.

Check out https://www.electric.ai/careers, and feel free to shoot me an email at nathan.gould@electric.ai.

Seconded

You comparing apples and oranges. Subsidies are a direct transfer of wealth, whereas Amazon's investments would remain assets of Amazon.

Electric AI | New York City (NYC) | Software engineers, Data analysts | Full-time (Onsite) Electric is the world's first all-in-one, real-time IT support solution for small and midsize offices. Through a chat interface, personalized service and flat-rate pricing we keep your email, computers, Wi-Fi and software running smoothly at a fraction of the cost while eliminating headaches normally experienced with traditional managed service providers.

Behind the scenes, we're building out a hybrid human/software platform to resolve and execute IT tasks with maximum efficiency and automation.

Electric AI is backed by Bessemer Venture Partners, Primary Ventures, and others. We're currently a few months post- Series A, and hiring across the board for engineering roles. Our stack: RoR/React/Redux running microservices on Heroku/AWS, plus some Python for data infrastructure, ML, etc.

Check out https://www.electric.ai/careers, and feel free to shoot me an email at nathan.gould@electric.ai.

Electric AI | New York City (NYC) | Software engineers | Full-time (Onsite)

Electric is the world's first all-in-one, real-time IT support solution for small and midsize offices. Through a chat interface, personalized service and flat-rate pricing we keep your email, computers, Wi-Fi and software running smoothly at a fraction of the cost while eliminating headaches normally experienced with traditional managed service providers.

Behind the scenes, we're building out a hybrid human/software platform to resolve and execute IT tasks with maximum efficiency and automation.

Electric AI is backed by Bessemer Venture Partners, Primary Ventures, and others. We're currently a few months post- Series A, and hiring across the board for engineering roles. Our stack: RoR/React/Redux running microservices on Heroku/AWS, plus some Python for data infrastructure, ML, etc.

Check out https://www.electric.ai/careers, and feel free to shoot me an email at nathan.gould@electric.ai.

Care to enlighten us pension finance non-experts on what the article misses?

This is the first time I've come across the approach of storing database snapshots and saving them in a data lake. Do you find those snapshots are useful/used for analytics or data science end-uses, or are they more there for debugging and answering one-off questions?

As a startup employee, I'd say a big con is not necessarily the amount of equity, but the type of equity, and difficulty of evaluating it. Common stock really only provides exposure to the tail end of good exit outcomes, since other entities on the cap table get liquidation preference, participation, and so on. To make it worse, it's not standard to share information about capitalization with employees, so we have no real shot at valuing the equity portion of an offer. It would be nice to be able to know what our equity is worth, and get some benefit from a decent but not great exit.

A more apt question might be, would you be willing to take a pay cut to achieve an increase in happiness and live wherever you want?

I like Altair, and have built a couple analyses and products with it. In general, it's a convenient way to do exploratory data analysis. But I have found it to be sort of inflexible for simple variations on common chart types -- especially if you want to augment a chart with information that is not in the dataframe instance used to instantiate the altair.Chart object. For instance, if you want to draw a line overlay on a chart, the best way to do it (to my knowledge) is append the start and end coordinates to your dataframe before calling mark_line().

I'm curious whether anyone else has had similar frustrations? I wonder what enhancements might make practical use of Altair for those use cases easier.

I think your argument is logically correct, but you are using numerical assumptions that are off by one or two orders of magnitude. Your typical successful algorithmic trader is probably flipping their metaphorical coin 1,000,000 times, and getting 520,000 heads. Each individual trade may only be slightly profitable, but there is often no statistical ambiguity about the effectiveness of the strategy.

Individual trading strategies often become less effective over time, though. Whether this kind of success can be sustained at the level of a trading firm over many years is an entirely different question. Whether they can beat the market after fees is a third, also entirely different question.

Thank you for bothering to explain this.

Agreed.

Congrats, this is awesome!

Not odd at all, actually. Numpy might implement certain functions differently from other libraries. With a background in algorithms, you could make an educated guess as to complexity, but without knowing the exact implementation it's still a guess.

This is a fairly brilliant referral advertising scheme.

In the U.S., the main reason prices go negative is because of the production tax credit. Wind producers get a subsidy per MWh produced. Since the unsubsidized marginal cost of production is zero, the result is a negative effective production cost. So if wind is the marginal energy producer, the clearing price goes negative.

I don't know how things work in Germany, but I'd be surprised if the negative prices there had much to do with the costs of starting/stopping conventional units. In most electricity markets, participants have to bid in their marginal cost. Even if you're an inflexible nuke, your marginal cost is positive. You usually need genuinely negative marginal cost units to drive the clearing price negative. That only really happens when subsidies are part of the picture.

A few other big ones IMO:

7. Energy density (as opposed to power density; i.e. c-rate matters)

8. Temperature sensitivity (many batteries lose voltage at low temps, e.g.)

9. Ability to hold a charge

This comment got my gears turning about good ways to split up Uber in an anti-trust case. One possibility is to have Uber A provide dispatch as a service, while Uber B collects fares, disburses driver payments, and provides the front-end UI. Uber B could also set up whatever driver incentives they wanted. Uber A would be a regulated monopoly, and would have to allow other companies to compete with Uber B on its platform.

In this scenario, dispatch would basically be a public utility, and drivers could trust the algorithm not to cheat. I.e. since Uber B and its competitors would be running any driver incentives, Uber A wouldn't even know about it. (And legally shall not know such things.)

I realize this is a half-baked pipe dream, but an interesting thought experiment.

Yes, it's worth investing the time. Grammar of graphics is life changing if data visualization is part of your day to day.

Gooulo, agreed insofar as we should reject hate in its various forms. It's ugly, terrifying, and has nothing to do with what has made this country what it is.

But even as a progressive democrat, I feel sympathy for djschnei's response here. I think we should embrace and celebrate free speech in all of its forms. And when racists speak publicly, we should publicly reject their views.

When you suggest that large swaths of people should be "systematically suppressed," I think you're driving a wedge between yourself and a lot of fairly moderate conservative folks who could be allies in protecting civil society from the likes of actual Nazi's, etc.

My understanding is the CUDA interface is for NVIDIA GPU's only. There are non-CUDA API's for other types of GPU's, but they don't necessarily work out of the box with neural network libraries like Tensorflow yet.

Cool!

Mortgage backed securities were not risky because they were diversified. They were risky because nobody fully understood how undiversified they were, especially at the lower grade tranches.

Meanwhile, LC was alive in '08, and you can look at how their notes performed. Investors lost single digit percentages. In other words, they beat S&P by a lot.