I really like the FDR fireside chat idea

10 days of darkness (after they shut down social media /end 230 exemption)

How to sell option?
Here let me explain through simple math.
Say BN at 30000
30000 straddle trading at combine premium 800 ( based on current volatility)
Put at 400
Call at 400
You don’t know direction. Sold both.

If move will not come than both will decrease gradually.

At day end it will come down to 720-740 if move will not come. Difference is your profit.

Now case 2:
Move started.
At one points combine premium will start increasing.
Say 810.
Exit immediately. No need to do adjustment.
Let market settle.

Sell again straddle once market settle.
Now repeat same thing.

Straddle selling webinar over.
Good luck.

One can keep SL 550 for both strike to escape from sudden movement.

(THREAD)

1. There's been a lot of news lately about people in the Autonomous Vehicle industry realizing the problem they're trying to solve is much trickier than they initially hoped.

I think there’s a good analogy in a similar technology: self-checkouts.

2. Self-checkouts have been around a while. Companies have been working on them since 1984, and they’ve been in stores around the country for nearly two decades. As you well know if you’ve used one, they also still kinda suck.

3. This is not for a lack of financial incentives! A company that could figure the technology out would be able to replace billions a year in labor. Even a marginal improvement on the tech would be incredibly valuable as current systems feature staggering losses to shoplifters

4. It’s not as if “scanning items at a checkout” is an especially daunting task. It’s considered unskilled labor. Teenagers do it.

5. Turns out that making an automated system that’s 95% as good as a human is relatively easy and one that’s 100% as good as a human is very hard. I think it’s becoming clear that autonomous vehicles are going to turn out like this

शमशान में जब महर्षि दधीचि के मांसपिंड का दाह संस्कार हो रहा था तो उनकी पत्नी अपने पति का वियोग सहन नहीं कर पायी और पास में ही स्थित विशाल पीपल वृक्ष के कोटर में अपने तीन वर्ष के बालक को रख के स्वयं चिता पे बैठ कर सती हो गयी ।इस प्रकार ऋषी दधीचि और उनकी पत्नी की मुक्ति हो गयी।

परन्तु पीपल के कोटर में रखा बालक भूख प्यास से तड़पने लगा। जब कुछ नहीं मिला तो वो कोटर में पड़े पीपल के गोदों (फल) को खाकर बड़ा होने लगा। कालान्तर में पीपल के फलों और पत्तों को खाकर बालक का जीवन किसी प्रकार सुरक्षित रहा।

एक दिन देवर्षि नारद वहां से गुजर रहे थे ।नारद ने पीपल के कोटर में बालक को देख कर उसका परिचय मांगा -
नारद बोले - बालक तुम कौन हो?
बालक - यही तो मैं भी जानना चहता हूँ ।
नारद - तुम्हारे जनक कौन हैं?
बालक - यही तो मैं भी जानना चाहता हूँ ।

तब नारद ने आँखें बन्द कर ध्यान लगाया ।

तत्पश्चात आश्चर्यचकित हो कर बालक को बताया कि 'हे बालक! तुम महान दानी महर्षि दधीचि के पुत्र हो । तुम्हारे पिता की अस्थियों का वज्रास्त्र बनाकर ही देवताओं ने असुरों पर विजय पायी थी।तुम्हारे पिता की मृत्यु मात्र 31 वर्ष की वय में ही हो गयी थी'।

बालक - मेरे पिता की अकाल मृत्यु का क्या कारण था?
नारद - तुम्हारे पिता पर शनिदेव की महादशा थी।
बालक - मेरे उपर आयी विपत्ति का कारण क्या था?
नारद - शनिदेव की महादशा।
इतना बताकर देवर्षि नारद ने पीपल के पत्तों और गोदों को खाकर बड़े हुए उस बालक का नाम पिप्पलाद रखा और उसे दीक्षित किया।

Most software founders, particularly in B2B, need to get radically better at sales. @Steli , who is the best person for explaining the scrappy stages in the beginning and then building out a team with scripts and processes, wrote a guide to:

For technical founders it is irrationally, obscenely hard to reverse years of programming (ba dum bum) that sales is a value-destroying activity. Sales is CLEARLY a value-creating activity, contingent on you have a value-creating product.

The world will not drop what they are doing to adopt your work. This is particularly true in B2B, where simply building a better mousetrap won't overcome the activation energy required to get people with additional non-mice problems to prioritize changing mousetraps today.

This is very non-obvious for founders because founders are not often people who *want* to be sold to. We often come from a background where trying out tools is a bit of a fun hobby. We like looking at all the options, making charts, and ripping out partially complete tests.

"This week I unsuccessfully trialed four software options for automating that thing that has been killing us. Our actual production process remains the same as last week. Don't worry; this was a great use of time." is not a thing you want to write in a progress report to manager.

Important paper from Google on large batch optimization. They do impressively careful experiments measuring # iterations needed to achieve target validation error at various batch sizes. The main "surprise" is the lack of surprises. [thread]

https://t.co/7QIx5CFdfJ

The paper is a good example of lots of elements of good experimental design. They validate their metric by showing lots of variants give consistent results. They tune hyperparamters separately for each condition, check that optimum isn't at the endpoints, and measure sensitivity.

They have separate experiments where the hold fixed # iterations and # epochs, which (as they explain) measure very different things. They avoid confounds, such as batch norm's artificial dependence between batch size and regularization strength.

When the experiments are done carefully enough, the results are remarkably consistent between different datasets and architectures. Qualitatively, MNIST behaves just like ImageNet.

Importantly, they don't find any evidence for a "sharp/flat optima" effect whereby better optimization leads to worse final results. They have a good discussion of experimental artifacts/confounds in past papers where such effects were reported.