btrettel

20 ms·

181.
by btrettel 4y ago
Possibly could be this: https://www.youtube.com/watch?v=EshxZVMURt4 https://popl18.sigplan.org/details/POPL-2018-papers/3/Formal... The city is Los Angeles, though the content doesn't se
182.
by btrettel 4y ago
I wrote an analyzer for patent claims: https://github.com/btrettel/plint There are tons of ambiguities and other problems in issued patents! When I get the time, I intend to rewrite a significant amount of the code in
183.
by btrettel 4y ago
That's a pet peeve of mine too. Here's a relevant HN comment I wrote about my experience as a patent examiner that applies more broadly: https://news.ycombinator.com/item?id=31315234 > There are way too many jo
184.
by btrettel 4y ago
Entropy is basically a useful quantity. It's no more mysterious than enthalpy or Gibbs free energy, both of which have also caused me confusion in the past. To me, the issue with entropy is that it's initially presented without cl
185.
by btrettel 4y ago
It took a couple years in college for me to understand entropy. Entropy in classical thermodynamics is presented in a mysterious way that leads to confusion. Entropy in statistical thermodynamics, however, is logical. Once one understands b
186.
by btrettel 4y ago
The author mentions the "diamond-in-the-rough" strategy used by organizations, which in my experience can be particularly disrespectful of the interviewee's time. One thing I try to figure out how before applying to a job is
187.
by btrettel 4y ago
As a cyclist, I used to talk to some drivers who did particularly dangerous maneuvers near me. Often I talked to them at stop lights, though not always. The goal was to get them to think about how what they did was dangerous and didn&#x27
188.
by btrettel 4y ago
Someone did something similar before, but unfortunately it's now offline. The HN link is still up, though: https://news.ycombinator.com/item?id=19694578
189.
by btrettel 4y ago
Do you have a link or reference for details on Paul Glover's "poison pills"? I couldn't find much with a quick Google search so I don't think I have the right keywords.
190.
by btrettel 4y ago
The USPTO has some training for registered patent attorneys and agents: https://www.uspto.gov/about-us/events/international-agents-a... It's apparently based on the training examiners get. The USPTO also post
191.
by btrettel 4y ago
You can submit prior art as a third-party while the application is being examined here: https://www.uspto.gov/patents/initiatives/third-party-preiss... A lot of people propose things similar to what you have, but
192.
by btrettel 4y ago
> In general it has, it is just that the USPTO's software hasn't. On the contrary, USPTO's internal search tool (PE2E Search) is overall good and has many features that I wish public search engines like Google had. The mai
193.
by btrettel 4y ago
See this other series of posts I made and the links from them: https://news.ycombinator.com/item?id=33506241 I don't see how GPT-3 could be particularly useful for me as it exists now. GPT-3 doesn't know anything
194.
by btrettel 4y ago
I love research, but I hated most of my PhD. "Publish or perish" attitudes caused a lot of friction for me. I try to be more careful in my research than most, and I don't care too much about the sheer number of papers I publi
195.
by btrettel 4y ago
Interesting. I stand corrected. > This comes up almost immediately when studying to be a parent examiner. To my knowledge, this wasn't part of my training. Perhaps you mean studying for the patent bar? I'm not a patent attorney
196.
by btrettel 4y ago
1. and 2. Depends heavily on the patent classification and seniority. I get around 20-25 hours for a first action, much less for subsequent actions. I'm a newbie, though. Primary examiners in my area probably get only about 10-13 hours
197.
by btrettel 4y ago
The situation here is different than in other government agencies. The USPTO is funded entirely by user fees, not taxes. In fact, the USPTO actually diverts a lot of money back to the federal government, so they don't even get to keep
198.
by btrettel 4y ago
> I feel bad for them myself, because their tooling is apparently terrible. Makes me wonder if I should take a look into the space and see if I can come up with some ideas for a halfway decent set of tools. The search tools are acceptabl
199.
by btrettel 4y ago
This question is hard to answer as the time given varies depending on the patent classification and seniority. For the application which I'll post an office action for this afternoon, I will be credited 22.6 hours. (I can't claim
200.
by btrettel 4y ago
I think AI searches could be much better even without an improvement in the technology. Probably the best AI patent search tool would be written by a ML engineer who has spent time examining a wide variety of patents. The current tools seem
201.
by btrettel 4y ago
I have no idea as these tools rarely explain how they work. There are a lot of AI patent search tools and I'm sure each of them works differently. Also, don't underestimate how difficult finding synonyms is. Many examiners, myself
202.
by btrettel 4y ago
To be clear, I do think that patent searching has improved. My point is that it hasn't kept up with the shear number of documents available. You have a point about the switch to computerized searches. Searching by examiners used to be
203.
by btrettel 4y ago
> I have serious doubts that more examination time is the solution here - at the end of the day, the only people that have an incentive for completely thorough search are the people filing or defending themselves against the patent. And
204.
by btrettel 4y ago
No, patent applications don't need to be filed via fax. I've heard bad things about the various USPTO websites, so fax might have been the guy's preference as I can see it being relatively simpler. (Again, like my other comme
205.
by btrettel 4y ago
I'll agree that many of the tools are bad. I have a particular dislike for how slow Word is. But, I don't think that transitioning everything to the cloud is necessarily going to help. Word is slow many times because it freezes
206.
by btrettel 4y ago
I've thought exactly the same thing before. In my view, the amount of time an examiner gets should be directly tied to the amount of prior art needed to search. And fees should automatically increase as the amount of prior art increase
207.
by btrettel 4y ago
The main problem is searching for prior art. There are AI search tools but they usually aren't good in my experience. It's rare that an AI search tool will return prior art that can be used in a rejection. Usually the prior art th
208.
by btrettel 4y ago
Current USPTO patent examiner here. The most effective way to eliminate bad patents would be to give examiners more time, say double the time across the board to start. If an examiner can't find prior art in the little time they'r
209.
by btrettel 4y ago
In my view, Fortran's biggest advantage over C and C++ is its much simpler mental model. Compared against C and C++, a Fortran programmer needs to think much less about pointers and memory addresses. For example, Fortran procedures (fu
210.
by btrettel 4y ago
> This is true if you need large amounts of resources ($) to achieve your research goals. My experience is that the quoted situation is the exception in US academia. Despite paying a pittance, graduate students are expensive enough to ma

More ›