Hacker News new | comments | apply | show | ask | jobs | submit | sbisker's comments login

I did make you a beautiful logo, didn't I ;) Yes, it's been a while.

One of the reasons I signed and took your $25 offer was because I wanted to live in a future where I didn't feel massively entitled to someone else's successes. At some point it makes sense to lock in your friendship with your friend-who-you-decided-to-maybe-cofound-with-but-stayed-in-Boston-instead and move on with your life, before money gets in the way.

That said, every story is different. Just because all I made for borski was a silly logo, doesn't mean that this Cruise fellow didn't make a more significant contribution. For better or worse, it looks like that story will play out rather publicly now.

-----


I do kinda miss the rivets ;)

It will be very interesting to watch this play out.

-----


Once it becomes straightforward to do ML with a pre-packaged library, you'll quickly start to see Amazon or a third-party offer effective MLaaS. That will suck quite a bit of oxygen out of the room.

-----


Already exists:

https://aws.amazon.com/machine-learning/

Problem is, writing an effective machine-learning model already doesn't require knowing the algorithms well. It requires knowing your data well. You can provide tools for this, and AML does, but there's no substitute for actually working with the data day-in-and-day-out and developing an intuition for it.

(Deep learning promises to change that a bit, since the relevant features are extracted for you by the algorithm and you don't need to do any particular data cleaning or feature extraction work. You still need to understand your data well to understand how to train the model, though, and how to apply primitive ML operations - classification, regression, clustering, etc. - to a real-world problem.)

-----


You still need to understand your data well to understand how to train the model, though, and how to apply primitive ML operations - classification, regression, clustering, etc. - to a real-world problem.)

And this is the kind of stuff that I believe can be done by people who don't necessarily need phd's in Stats or ML. A decent grounding in statistics / ML, and good domain knowledge should be enough to support using pre-packaged algorithms to solve business problems.

-----


This is just a long-tail SEO play, nothing particular to Google Fiber. See also http://www.xfinity.com/compare/comcast-xfinity-vs-att-u-vers..., http://www.xfinity.com/compare/, etc. I'm guessing people googling "xfinity vs <competition>" is a phrase they'd like to bring in more traffic for.

-----


Theranos' reply: http://fortune.com/2015/12/20/letter-to-the-editor-theranos-...

It reads like it was written by a small team of lawyers who have had little sleep - not a good sign. :/

-----


Actually, I think it reads more like "we weren't deliberately misleading Roger Parloff, we were trying to mislead everybody!"

-----


The whole thing sounds so pendantic, like "See? We are TECHNICALLY CORRECT!"

-----


That link appears to have been deleted, so here it is, via Google cache:

----

Letter to the Editor: Theranos Responds

Brooke Buchanan is vice president of communications at Theranos

Both the headline and much of the content of Roger Parloff’s December 17 story, “How Theranos Misled Me” are inaccurate and — ironically — misleading. Theranos was honest and transparent with Mr. Parloff, and the headline to Mr. Parloff’s story is not supported by the facts or the story itself. Let’s take a look.

After two months of reviewing his notes, there is no statement by Theranos cited in this article that is in any way inaccurate or misleading to support his headline or the statement in his article that he was misled.

Rather, Mr. Parloff states in this story that he made an error: “I then started looking back at my research for the original story—which had been conducted, by then, 17-19 months earlier—to try to reconstruct how I made the error (italics added).” He concludes, “I regret the error.”

Mr. Parloff’s basis for saying he was misled appears to rest on the last part of the following statement in his lengthy original article: “I wrote that the company currently offers 200 – and is ramping up to offer more than 1000 – of the most commonly ordered blood diagnostic tests, all without the need of a syringe.”

The fact is that Theranos was, indeed, offering 200 diagnostic tests at that time, and was ramping up to offer more than 1000 – all via finger-stick. That is true.

Theranos had also developed over 200 assays to run with small volumes from finger-sticks, urine, and other sample types at that time, as we explained to him then. That capability remains today.

Theranos did in fact run certain tests collected through venipuncture on its proprietary technologies at the time of this article.

Theranos never told Mr. Parloff it was running all of its tests without a syringe – neither did he ask and neither was this a secret. Our website has always made clear that was not the case (e.g. “Instead of a huge needle, we can use a tiny finger stick or collect a micro-sample from a venous draw. Theranos website, 2014). Indeed, Theranos has always been clear – in statements on the website dating back to when it announced its retail lab services in 2013, in statements it made to those who asked, and elsewhere – that it has always performed some tests via venipuncture. As acknowledged in Mr. Parloff’s story, Theranos explicitly discussed its use of venipuncture with him.

To be clear, this topic was also not fact checked with Theranos prior to his article. Rather, as he readily states, this sentence was based on assumptions that led to a statement in his article which he is now reading through the lens of other’s faulty reporting – reporting based largely on anonymous and uninformed or misinformed sources with respect to our proprietary technologies and their capabilities.

In his article, Mr. Parloff now says that he feels he simply “assumed” that Theranos had brought up over 200 tests to run in its clinical lab only on finger-sticks. Again, Mr. Parloff never asked anyone at Theranos whether that assumption was correct. Had he done so, we would have corrected that ambiguity. And if the sentence had jumped out at the time as containing an erroneous assumption, Theranos would have corrected it.

Finally, it is important to point out that when Mr. Parloff first met with Theranos about writing the article, in the spring of 2014, it was originally in connection with a patent troll litigation that Theranos had undertaken. It was in that context — of demonstrating Theranos’ technological capabilities and other intellectual property — that Theranos shared data from the development of over 200 proprietary assays on finger-stick/small volumes of samples.

If Mr. Parloff now feels that because of recent reporting he got one sentence in his story from mid-2014 about Theranos wrong, it was not because he was misled, but rather because — as he states — he now feels he made faulty assumptions that he now thinks he should have further clarified. The company has grown and evolved and made many business and regulatory decisions about its operations along the way. At no point did the company mislead him.

-----


Still nothing here about how many of these tests were actually reliable. I have a 100 fairies at the bottom of my garden who can tell me the result of any test without even drawing a drop of blood. Profit!

-----


Heh... Akin to how the founder of the mormon church came up with the idea while getting high and sticking his face into a magic hat. Get loaded, stick your face in a hat, document your delusional visions & pander them upon the easily swayed ... voilà! You have created a $7 billion a year religion!

-----


Still there for me.

-----


You bet they have! The reputational risk to a media outlet of writing a digital puff piece is now negligible; they can basically drive traffic to or away from it as they please. There are also real SEO benefits to mediocre pieces beyond ordinary daily readership.

-----


We've seen a rash of these "bad news bears" articles lately. It's important to note that this stuff has always existed in some startups, and it isn't inherently bad - it's a natural consequence of taking so many risks on inexperienced founders and managers. Some will train a new management elite, some won't quite grow into their expectations and some will devolve into Lord of the Flies.

It's just that as the tech market becomes softer, people stop deciding "my stock will go up" is a good reason to put up with terrible management. The hard part for people in startups is having the courage to walk away when things are still growing without a care.

-----


The fact Fidelity is being this transparent is great news for engineers in the startup scene considering or holding jobs at these companies.

No longer can managers keep selling a dream that was last pitched years ago at their last funding round; they'll have to compensate for underwater options with rising salaries or grants to keep talent around. Certain startups have a habit of delaying sharing bad news with their employees to the very brink of collapse; this behavior will now be in check somewhat.

-----


I heard that. Employees need more data to evaluate job offers (especially in this economy where the big guys are offering a lot of cash, and the small guys are offering a lot of cash too but also equity of sometimes dubious value).

If you take equity, you are not only subject to dilutions, but also to other investors getting much more preferential terms to you.

I'm not saying to don't take a job because of shares, RSUs, or options offered, but one needs to take a hard look at these things to figure out what their true worth is. And in most cases, you may only be able to compute a confidence interval and not a point estimate.

Fidelity's disclosure makes this seemingly intractable problem a bit less intractable.

-----


Curious to see how much devaluations push salaries across the board up, rapidly increasing burn rate of on-hand cash.

-----


So by publicly devaluing a portfolio company, Fidelity increases that company's hiring costs. How is that in Fidelity's best interest? Sure, they're devaluing the company, but they still have a stake in it... why be so public about the devaluation?

-----


I imagine Fidelity's fund customers would want to know.

-----


Fiduciary Duty.

-----


Which is far more important (from a liability standpoint) than pissing off companies you're invested in.

-----


It's not at all clear they had a solid support and sales setup. The company folded for a reason - why would you buy a company and run it the same way it was run before, if it was about to fold under those same processes?

-----


Yeah, that's a trick drivers play on customers so they don't take a hit to their rating. If Uber catches a driver canceling too often, they'll boot them or freeze them from picking up more fares. (Conversely, if a customer is doing the same thing too often, a similar penalty of fees or the like will happen to them as well.)

So next time they try that, feel free to hold your ground by using Lyft instead or something. You don't have to cancel - they must, by their own rules, either cancel themselves, or pick you up.

-----


I suspect the list is pretty long. But off the top of my head, Pandora had a pretty brutal layoff in the very beginning, to the tune of (from what I understand) basically asking their most dedicated employees to work for free. Somehow they pulled themselves out of that one, but, it wasn't pretty.

-----

More

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | DMCA | Apply to YC | Contact

Search: