mmmBacon

Direct link to the job in question:

https://job-boards.greenhouse.io/anthropic/jobs/5285248008

Wow this is awesome! I’ve been using ChatGPT to create TikZ diagrams because honestly it’s not trivial to create TikZ diagrams and they look really good when done well. However it can be challenging to get ChatGPT to tweak the diagrams. My knowledge of TikZ syntax and behavior isn’t always sophisticated enough to get the outcome I want. So this thing looks like an answer for that. Looking forward to trying this.

Incidentally I am using ChatGPT to create Latex documents for my blog and then they get rendered server side. The workflow works really well.

UofR physic grad that also worked at the LLE here. Agree Rochester schools are underrated (although admittedly a little biased).

At least in the sciences you have access to lots of opportunities you don’t have at bigger name schools.

They set me up in life in a way that I don’t think would have happened elsewhere.

It’s ironic what you wrote because it was Carl Sagan who first developed the theory to explain the Venusian climate.

I met Carl Sagan and he was not a person who dumbed down anything. He had a profound impact on planetary science on top of inspiring many including myself to pursue physics.

I think the fact that we are all still talking about Carl 30 years after his death is strong evidence of his impact.

We are still in a regime where the growth of logical qubits is far below the point of any practical computing utility let alone at a level capable of breaking existing codes. While it’s correct to be proactive at the security end, if you take the current rate of improvement in logical qubits and project it forward you’ll get something beyond 2040. I built a probabilistic model of logical qubit improvement and the median ends up being further out (probably should publish this somewhere). I’d like to point out that I’m not purely negative; I was surprised to calculate we can make enough He-II to make it all go!

Of course such a model cannot predict a fundamental breakthrough nor can it predict whether there is some kind of fundamental limit to the size of such a quantum system before we have coherence collapse. This is an interesting question for quantum mechanics however.

In summary, quantum computing feels analogous to fusion, a technology that’s always 20 years away.

Oh and don’t get me started on AGI. Lol.

Totally. I started baking pies because it was a tradition in my family and my wife can’t cook. To make sure my kids had the family food tradition I learned to bake. Once you get a system down, like anything else, it’s not that hard. Plus pie filling has time to bloom if you make it day before. Pie dough can be made ahead and freezes well. Individually these things aren’t hard or time consuming.

I started making my own simple bread and now I can’t eat store bought bread. Just takes like sawdust to me. It’s not really all that hard. Add a little rosemary and some olive oil and it’s delicious. No need to fuss over sourdough (over rated in my opinion). Over time you learn how ingredients work and what ratios work. So becomes easier and easier. I can throw together amazing corn bread and be eating it a little more than half hour later.

The GP has a point about the state of journalism generally and the pervasive nature in which Yellow journalism is returning.

One need not be anti-intellectual to find the state of reporting to be difficult to deal with and not wanting to read it. In addition to the GP’s complaint; journalists of any ilk also tend to conflate editorializing with reporting. You see this all the way from pop science to NYTimes to Fox News and yes even the Economist.

A question is whether the more fact based reporting of the early-mid 20th Century is the exception to the tendency of Yellow journalism that existed before and seems to exist now.

Midnight PRs sounds really sad. Pager duty; I mean you’re not saving people in the ER. Everything you’ve written comes off as profoundly selfish and self-centered. God forbid your 8 month old needs to be a priority over pager duty.

Raising kids is hard, I have 3 but it’s not sad. Blowing off some steam is something every parent needs. But it sounds like you are in desperate need of some perspective on life.

By what metric is the US lagging? By any objective measure we can see the dominance of US technology. I think it’s most of the rest of the world that’s being left behind; Europe in particular. If what you’re saying is true the US economy would also be flagging but it’s not. If what you’re saying is true, you’d see the list of the world’s most valuable companies dominated by non-US firms.

I think you are confusing the current climate of immigration enforcement and reform with being anti-immigration. The US will continue to draw top talent because the US is where the bulk of the opportunities are and will be for at least the next 5 years.

It’s been widely discussed that the immigration system has been abused, especially by the tech industry. This reform started under Obama. The current outcry is a reaction to the most recent federal election. Reform does not mean the US is anti-immigrant. It may mean lower levels of immigration that’s more selective for talent.

Parental controls are all about illusion of control. In reality the kids all know how to get around them. Try just try to block Gmail. It’s impossible. Gmail is the gateway to kids getting on services they aren’t supposed to be on.

Gmail can circumvent almost any security feature even if you set up a profile on iPhone (which is not documented and good luck with that). This is definitely not an accident.

Don’t mean to pick on Google; Apple is also bad, iPhone parental controls are very leaky. My son found a way to jailbreak his phone to completely unlock screen time and disable all parental controls.

Any of the consoles are also bad, PS4, etc… although it is possible to block stuff that PS4 can do via a firewall.

Why do you assume I don’t think that? Of course it will happen at some stage. But my comment wasn’t addressing the future, but now. Effectively ChatGPT is buying market share by not having ads; that’s clear. What’s not clear is whether there’s a more innovative model than inserting low-quality, mostly irrelevant ads into the body of the chat they way Google and YouTube do it today. In a Chat agent, advertising also will be an issue for credibility of results. So I think the ad model as we know it will need to change.

I think Google’s search and ad business are at risk. Search has become such a mess that it’s become harder and harder to use to find quality results. It reminds me of Yahoo before Google in a way.

I’m using ChatCPT or equivalent for 60% of my searches. The remaining 40% is just muscle memory. Of that 40% about half the time I regret using Google search due to the difficulty of finding the relevant result.

I can see search users moving to ChatGPT or such and Googles Ad business suffering as a result and a general downward spiral of Google search.

Sure, by the time China clones this generation of tin droplet ASML EUV machines at production scale, the market will have shifted to free-electron lasers.

Yep you provide a great example of a word used as regional slang for the word television.

The word telly is not in common usage in the United States. It’s understood here to be UK slang for TV.

Your example is largely irrelevant; I wouldn’t call a spyware TV founded by Russian born dude a cultural touchstone.

Regardless of origin the word television is an English word now. The ability to adopt loan words from other languages is one of the many reasons English usage is so widespread.

Isn’t “telly” English slang for television? It’s a regional slang that’s not universally used.

This is 100% incorrect as you’ve written it. The GRE is based on English vocabulary. It’s true that many words have Greek, Latin, or French roots but they are most certainly not Latin, Greek, or French.

I’m old enough to remember being able to scrounge around the house for pennies and heading down to Gracie’s corner store so I could buy some Swedish fish. They were 1 cent each. Gracie counted them out and put them in a small paper bag for you.

A major score was finding a dime or quarter on the street. When the Whatchamacallit first came out they were 25 cents!

Maybe the jury needed 7 hours to determine if the bread was stale enough to cause bodily harm. Perhaps the crime here is the waste of a perfectly good sandwich.

Visible will always be expensive because it’s very niche and low volume. So the techniques here are only practical economically for the large volumes of light sources required for communications. This won’t extend to the visible unless there’s a similarly large market.

The cheapest way I’d think to generate a visible frequency comb would be to frequency double the IR comb laser using a nonlinear crystal like BBO.

Also here the accuracy is relative and not absolute which is fine for communications. The absolute accuracy of the comb may not good enough for spectroscopy in the visible.

Not surprising at all to me after several summer trips to interior Alaska. The mosquitoes are so thick that you inhale them sometimes; which is so disgusting. I slathered myself in Deet (the only thing that works) and was mostly ok. Even then they find every square mm that you missed. I sat down for 30 minutes on a bench leaning forward talking to some people. My shirt pulled up about 1/2” (12mm). Later I counted 137 bites (some had merged due to swelling) across that strip of exposed flesh!

Concur your response; you can get a 48hr transit visa on demand in China. The requirement is that you leave via the same port of entry.

Do you have any data to back this claim up or are you just stating your opinion?

I was routinely detained at passport control because there was a bad guy with my same name. It took some amount of time and being very polite to get me out of that.

I run a team. So for me this is like having 2-3 extra people to just do some exploratory work and prototyping. Having actual people to do this is a luxury I don’t have and can’t justify currently.

I can’t just copy-paste what comes out. But I have to say I’m able to get substantially more done as it saves me a lot of grunt work. You do have to learn how to use it like any other tool. I have found that it’s helped me sharpen a lot of my own skills in multiple areas and improved my understanding of systems I’m working on. I am able to learn new things much faster because it has a real-time feedback mechanism.

I can’t speak to your direct report but I’d be concerned about falling behind as a leader if you aren’t using it in the course of your regular work.

I disagree. Couldn’t do what I’m doing with GPT-4 but I can with 5.

I am personally paying $200/month for the ChatGPT pro version. It’s the best investment I’ve made in my career in a long time. It’s like having 2-3 direct reports. The alternatives provided by my companies are too slow and aren’t anywhere as good. If people start using it anywhere near how I’m using it, there will be profound changes to the workplace in the coming 3 years.

The Surfliner in Southern California is a great deal. Take the business class ticket. OC to San Diego is a breeze and you can relax and enjoy all beach towns with a glass of wine. Took it from OC to Santa Barbara. It takes about an hour longer than driving but so worth it. Can’t say enough good things about it. Views in all directions totally amazing. Trains are packed at times but it’s good to see.

I think high speed cross country rail travel is not very practical either to build or for travelers. The distances are too large, terrain too difficult and population densities are too low west of the Mississippi. It’s worth noting that the US train network is optimized for freight and we carry much more freight than many countries by rail.

High population density corridors like BOS-NYC-PHL-DC and SB-LA-OC-SD make a lot of sense and would be great to see even better service. We are seeing a high-speed, private service building track between Ontario, CA and Las Vegas. Distances to Vegas and Phoenix are short enough and population densities high enough that it makes travel by train from SoCal competitive with driving and air.

Back in the 2000’s we built a complete network operating system on top of QNX for OTN based long haul communication systems. In those days we had to sign SLAs on equipment and customers fined us for downtime. QNX was bulletproof despite running on our then PowerPC based custom CPU complex.

This is cool and it amazes me how the “big” home security companies don’t have this. My Ring cameras false detect all the time. The front will detect the flag as a person when it blows in the wind. The rear gets triggered by the pool robot skimmer. As much as I’d love to try this, I don’t have the time. It would be great though if it was built into something I could buy.

If you think of it, what is Meta going to do with AI? Their business model is ads. Maybe AI can improve ads some. For Meta, how does the billions spent on AI impact their bottom line? The reality is that if AI training requirements keep scaling based on Chinchilla scaling, there is going to be massive consolidation in training due to the scale that’s going to be required. Not everyone is going to be comfortable with Zuck running nuclear reactors. Already AI models are the fastest depreciating assets ever created. Moreover Meta is behind the AI curve so they are desperate to catch up. But desperate for what?

The supply chain is already dispersed, even outside Taiwan. Particularly as we move from single die devices to MCM, many processes are outside of Taiwan. JCET is in Singapore and Amkor is in Arizona and Korea for example. There is some cost to the logistics but it’s kind of in the noise on a per device basis. The cost is in the processes themselves. It is a gigantic pain to manage but it doesn’t not add such a high variable cost.

Semiconductor companies need gross margins of around 65% to grow and be able to invest in development of the next node. So this large additional variable cost really can’t be shrugged off as you suggest. If so, Ms. Su wouldn’t have mentioned it at all.