rubidium

Yea I’m looking local hosted transcription and translation with diarization of 2 (or more ) speakers. This is to speed up collaborative technical work between two teams who speak different languages where want all local processing (assume no cloud access).

Any that also support translation? How much harder/ easier of a problem is local translation compared to transcription?

99% of this is the printhead and the ink formulation. Assuming you use generic off the shelf solutions for those two components you’re set. All the printer companies do their lock in at the firmware and software layer.

I did when I was in academia. Would open each day and check what new papers were in my field. It was fun, and I learned a ton.

I kept it up out of habit for a year after grad school. Then moved on.

And the biggest thing is that: software the way we want is much easier. No ads. No monthly cost.

You’ve not accounted for the danger of necrotizing bureaucratic systems. Put these LLM drones in enough places and everything wills stagnate.

It’s effectively the start to Asimovs Foundation.

I do agree my skepticism level rises extremely high in any experimental psychology experiment. There’s just so many ways to bias results, in addition to “do enough experiments and one of them will get a statistically unlikely result” problem.

This group does a lot like this https://www.dpg.unipd.it/en/compcog/publications … so that’s tempting to think they keep trying things until something odd happens (kind of like physicists who look for 5th forces… eventually they find something odd but often it’s just an experimental issue they need to understand further).

I did daily (old station wagon in the rear facing seat), as well as school buses. Kindof liked the smell in moderation as a kid.

Still in favor of EVs, just a curiosity that this is so negative for you.

It’s great place for grad school. Very little to do other than study and lab work. Play soccer with the grad students, float on the river in the summer, shoot potato gun cannons in corn fields, but otherwise there’s nothing better to do than work on a PhD.

I’d recommend it for grad school. But then yea time to move on.

This is the attitude of someone who uses hand tools when power tools are available. Yes you loose the personal touch but also loose the potential efficiency. Still need to measure twice and cut once though.

Make a list of the most interesting companies / areas you want to work in from a scientific perspective

Cold call or get introductions to their R&D leaders (principal researcher /director / VP). While connecting to something they or their company did, ask for a coffee meeting or phone call to learn more about their company and how it works.

Use this to Network network network. At some point a job will appear with interviews. Chances are good it’s with a good manager as they’re the ones taking time to build good teams and talent.

Amazon prime video showed us that.

First no ads. Then ad free if you pay extra. Then “ad free” except half the shows have a “this show requires ads” bs and still have ads. Scummy flea ridden advertisers at their core.

You went to the wrong university :)

It’s pretty much always who you know… at least to get a showing. It’s rare in history to find counter examples. And in a LLM fueled world it’s going to be more important.

Companies can improve by ensuring they don’t hire _because_ of whom someone knows. It should only ever let you get in the room to interview.

So practical advice of what to do: be human. Get to know people. Care. Your time to do this is not when you’re looking for a job, but when you’re in a job.

You have experience in 2 very different places demographically. High altitude may be on the list of factors but I would guess it’s pretty far down.

Boulder is a small college town of mostly affluent younger people. Memphis is an urban city in a hot climate that is older, poorer, and in the middle of the section of the US that has some of the worst health overall.

I’ll be concerned when ex-Yelp “growth strategists” start showing up at OpenAI and leverage the same extortionist technics.

This is an engineering management hot take on how to motivate a team. It’s not realistically covering burnout, human psychology, or what makes us human and tick. It’s just giving management advice with a flashy title.

It’s not necessarily bad advice (setting and celebrating milestone goals on a long project is good). But it’s also not applicable to most cases of disengagement from the important things in life (aka burnout).

When you become a good enough IC, “ As an engineer, I prefer to be managed and guided by someone who actually knows what I work on, preferably better than I know it.” is no longer reasonable. Then your managers role is to maximize your ability to make an impact by putting you in the right place/project.

As a manager of people who know far more about the things they do than I do, my goal is to assist and ensure in the right place (for them and the org). It’d be foolish for me to hire peons who know less than me.

I used them in graduate physics courses 2008-2010

For people I talk, it is the end of ZIRP that caused it. ZIRP ended in April 2022.

Even as interest rates went up the VCs still had committed funds to distribute for a year to 18 months. Then the biotechs had runway for 1-2 years from that. Now that’s all gone, and they can’t raise their next series. In the meantime, C&GT plus synbio is recently having a lower win rate than hoped for.

That plus all the money that is there is all going to AI companies due to the shorter time to return / higher potential roic / hype.

Biotech is facing a huge downturn right now too.

Fujifilm succeeded at it (before Kodak tried). Interesting case study between the two companies.

The overlap of statisticians and baseball fans is high (anecdotal).

Hypothesis 1: statisticians love good data sources, and with its many games, innings, and types of hits / pitches it’s a great source.

Hypothesis 2: makes you seem more interesting at dinner parties

hypothesis 3: a natural overlap of preferences

I have no idea why you’re using icons for 13 things. Use words.

They make it seem like a big deal. It’s pretty much how all software used to ship :)

You’ve taken a very small slice of the total workflow to develop cas9 crispr edits. Yes some components are “not crazy expensive”, but that’s not therapeutic production levels. And no sequencing. And no bioinformatics. And then zero tissue targeting/ delivery or safety studies.

So $5k is a massive stretch. I do think costs are coming down. But don’t want to oversell current state of “garage hacking” either.

I thought it supports most file storage sync solutions? I’m using with the free iCloud level storage I believe.

Yes. Aldevron and IDT are two companies (owned by danaher) that collaborated to make this happen. They have multiple authors listed in the NEJM article.

https://jobs.danaher.com/global/en/search-results?keywords=S...

More or less, yes.

An interesting part of the study was determining what a clinical dose _should_ be. You need enough to edit enough liver cells. But don’t really want to completely overdo it to limit potentially negative side effects. Seems like they got it right enough here, with the first dose having some effect and the subsequent dose having more.

It can make “bad edits” eg off target effects. But in this case there were, as far as is known, none. It’s aided that this was a single nucleotide defect.

They specifically tested for off target edits in the mouse study and found no harmful edits (and very rare off target ones). That plus the specific targeting of the liver cells (no germ line effect expected), makes this a low risk approach and certainly better than doing nothing.