Your therapist is the only person you know with no agenda
232 karma · joined September 6, 2013
Your therapist is the only person you know with no agenda
According to the article it sounds like they had ~14,000 employees
Edit: I misread. 20% of corporate jobs were cut which contributed to the 2800 number
It helps explain all aspects of the drug trade through the lens of business.
One of the most fascinating takeaways was how they handle “contract enforcement” in the absence of a legal system or courts (hint: Violence)
I recall an anecdote that Mexican cartels would hire Mexicans —- and not Dutch —- in order to serve as drug mules smuggling product into Amsterdam. Dutch mules would get caught less often, but Mexicans were much less likely to steal the drugs entirely. Because when you sign up to be a mule they take down the names and addresses of your entire family (nearby), and will kill them if you steal the product. Contract enforcement.
Being a Computer Science grad / software engineer and on the quieter side, I always labeled myself as an introvert. But I more recently identified that I get a lot of energy from talking to the right people, even if I’m not loud and gregarious like your textbook extrovert.
The main thing is I only get energy from talking to interesting people about interesting things, and I now see that as a positive, instead of hating myself for not meeting some unrealistic standard of being able to engage with boring people over small talk or trivial matters
It is definitely challenging to pivot and fighting sunk cost bias is massively hard
We are on a mission to make the physical world shoppable. We are using QR codes, NFC, Apple App Clips and Android Instant Apps to make it easy to reorder physical goods (like drinks, skincare, etc) and also to make it easy to buy merchandise at concerts and events.
Our focus is on creating immersive experiences that feel much different than a typical QR code or website. Check out an example by scanning the QR code on our website: https://getbatch.com
We have a small, senior engineering team of 4 and are looking for senior engineers to join the team. We believe in running a small autonomous team where each member is invested in the product and deeply trusted to build the right solutions for the business.
We just raised $5M from a series of top tier investors including Coatue and 776. [1]
Our stack is: Ruby, Rails, Heroku, Postgres, Redis, AWS, Lambda, React, Netlify, GraphQL, Apollo, Swift, Kotlin
Learn more and apply here: https://jobs.wrkhq.com/batch
[1] https://techcrunch.com/2021/06/23/this-qr-code-startup-just-...
https://charliereese.ca/article/top-50-y-combinator-tech-sta...
We are on a mission to make the physical world shoppable. We are using QR codes, NFC, and Apple App Clips to make it easy to reorder physical goods (like drinks, skincare, etc) and also to make it easy to buy merchandise at concerts and events.
Our focus is on creating immersive experiences that feel much different than a typical QR code or website. Check out an example by scanning the QR code on our website: https://getbatch.com
We have a small, senior engineering team of 4 and are looking for senior engineers to join the team. We believe in running a small autonomous team where each member is invested in the product and deeply trusted to build the right solutions for the business.
We just raised $5M from a series of top tier investors including Coatue and 776. [1]
Our stack is: Ruby, Rails, Heroku, Postgres, Redis, AWS, Lambda, React, Netlify, GraphQL, Apollo, Swift, Kotlin
Learn more and apply here: https://jobs.wrkhq.com/batch
[1] https://techcrunch.com/2021/06/23/this-qr-code-startup-just-...
They use a VPN config on iOS to monitor traffic, and on MacOS make their monitoring app very difficult to uninstall without an uninstall key (which notifies your accountability partners)
They're the same quality as Herman Miller but quite a lot cheaper
(disclosure: I'm an investor)
If a project is only going to ever use http, why would I need to bundle in all of the others?
Genuinely curious what the advantages are
Like when instagram started they promised they'd have no ads
What data are they observing from a telescope that tells them it’s iron and nickel, vs some other metal, vs rock? Especially with such precision around the composition
My background is as an application developer (RoR, React, Python) and fell in love with machine learning. I wanted to use ML at the startups I worked at but with everything involved it would become a bigger project than anticipated and fall out of priority.
The aim with Deepserve.ai is to make shorten the cycle enough that you could train and deploy a prototype model in a day and a production ready model in a week (if you had the data ready).
I am actually still part time and bootstrapped — I spend a few hours a day hacking on the platform and am starting to market to ML practitioners and sales into enterprise. It's definitely a lot of context switching — from React to CLI writing to AWS devops to marketing.
My aim is to be profitable out of the gate by having pricing with sufficient margin over my underlying compute cost. I'll definitely be more expensive than doing it yourself — but the point is that you don't have to do it.
If you launch your startup, definitely reach out, I'd love to add AI to your product! And if you want any advice on your startup, I'd love to connect, this is my third startup and I've had a couple of successful exits so might have some ideas!
My goal with Deepserve.ai is to be drastically easier than the big cloud providers, with less lock-in. Rather than giving you infinite configurability with low-level primitives, I want to give you an easy, scalable solution right out of the box. Sagemaker and Tensorflow Serving have some quick-to-deploy options like this, but are really built in expecting you to use their entire ecosystem.
I started off with Fast.ai for simplicity and want to expand to support PyTorch, Tensorflow and ApacheMX. My aim is to have an option that is tested and proven for every library version and every mix of inputs/outputs. So if you have a PyTorch 1.5.0 model that expects an image and returns a set of bounding boxes, you choose that configuration and have everything ready to go.
You're right on with having things like logging and tuning built in for everyone from the get go.
I will obviously never be cheaper than the big cloud providers. I want to compete on ease of use and the soft costs of hiring for or dealing with devops.
Thanks for your kind words!
We’re now open for beta! I want to onboard companies slowly as it’s critical our infrastructure is robust. Out of the gate I am supporting Fast.ai and working hard to support PyTorch and Tensorflow.
If you don’t have the skillset right now to build your own models — reach out on https://www.deepserve.ai/consultancy or email me. I can build models for you and host them Deepserve. I’ll be transparent that I want to make money on hosting, and would love to take your use case and turn it into a case study.
As an example — I have a client on the platform right now using a text classifier that auto-tags and auto-categorizes their support emails. These emails then get routed and sorted automatically, with the most urgent cases alerting them via slack.
—
This has been a side project for over a year, and am finally ready for launch. Feedback is welcome and I’d love to connect with anyone who aligns with the mission of bringing machine learning to the world.
You can reach me at jeff @ deepserve.ai
I’m Jeff Wells. I am launching https://www.deepserve.ai — a platform to deploy and host machine learning models.
Deepserve makes it extremely easy to take a trained model and deploy it with a single CLI command: `deepserve deploy`
We host the model, deploy it, and give you an API endpoint you can call on to make predictions from your applications. We manage the devops and dependencies and ensure your model is always running. We’ll scale up as much as your application needs, and scale down during off-peak hours to save you money over renting your own servers.
In order to make it easier to use, we have client SDKs to make calling your model a single line of code. We store all of the production inference data so you can grow your training sets with real examples.
You can run `deepserve deploy` on the command line and it will stand up a REST endpoint for your model. I'm also building client side SDKs to make it easier to call on these models in your application code.
It's still in beta but feel free to reach out to me jeff @ deepserve.ai
It was originally called Astra, which ended up having even more conflicts.
I use f.lux and love it. Hoping mine is different enough to not confuse. Funny enough, I actually thought it was “f dot lux” and was conceptual thinking of it like “function dot light”
Instead of renting a GPU instance and setting up a Flask web server, you use git to push your trained model to Flux with some configuration and get back an http endpoint.
For example, you set that your input is the url to an image, and that your output should be the top classification and its likelihood, and that your model is in pytorch.
For example if you have a classifier for dog breeds you:
Make a POST to fluxdeploy.com/username/dog-classifier with json { “url”: “...” } And get back { “klass”: “Great Dane”, “probability”: 0.937373 }
No need to do your own devops, Flux will scale for you. And it’s priced per-request and cheaper than hosting your own web server. Flux also deals with versioning and dependencies.
Still working on streaming inputs like video.
We were on Vendor Central so we had a dedicated rep. He would avoid helping us however, and use the opportunity to force us to renegotiate our rates (so Amazon could have more margin). Then the issue would disappear.
This happened so often we suspected these automatic issues were actually orchestrated to secure more margin.
The Amazon rep told us that he had quotas to hit of new margin targets for us every quarter.
We also consistently had issues of Amazon reporting under-deliveries and holding our funds. We shipped product in packages of (modulo) 8, and so they ordered in packages of (modulo) 8. When it arrived at the warehouse, they'd mark only 7 delivered, which is not feasible, sell the extra one and refuse to pay us for that 8th unit. This amounted to millions of dollars.
I eventually tracked every single package we had ever shipped and presented evidence of this fraud to them. They offered to pay us 30 cents on the dollar if we signed a waiver forgiving the rest.
If that person ends up duping you, then you need to sue them (litigation)
I have been doing sole crushing work for years in school. When you don't have a choice, the most useful thing for me to get started is the pomodoro method. Spend just 25 minutes of agonizing work and plan what you want to do for the 5 minute break. Usually after 1 or 2 cycles I actually get focused and motivated enough to make some progress.
I only looked at Ruby Basics, but here's my feedback:
The challenge is to solve: 'true and 0 && !nil and 3 > 2', but this is not a real-life application.
The words 'and' and 'or' should not be used for boolean comparisons, they are intended to be used to join two separate clauses such as "render 'template' and return" in rails.
Comparing a number and bang-nil is not good code. This could actually come to be if you were doing something like: "object.value && !object.value". However, you wouldn't ever want to hardcode "0 && !nil". If you were really checking for nil, you should use "!object.nil?"
I'm afraid that these challenges will teach bad code techniques to newbies who don't know the difference. But again, as someone with more familiarity with Ruby, I think these types of challenges are fun, but if that is the goal there is no need for the tutorial beforehand.