How do we fundamentally discover new things? In a letter to Maurice Solovine, Albert Einstein conceptualized discovery as a cyclical process involving an intuitive 'jump' from sensory experience to axioms, followed by logical deduction. While Generative AI has mastered Induction (statistical pattern matching) and is rapidly conquering Deduction (formal proof), we argue it lacks the mechanism for Abduction—the generation of novel explanatory hypotheses. Using Einstein’s formulation of General Relativity as a computational case study, we demonstrate that the prevailing theory of "creativity as data compression" (induction) fails to account for discoveries where observational data is scarce. This position paper argues that while a modern Large Language Model could plausibly execute the deductive phase of proving theorems from established premises, it is structurally incapable of the abductive 'Jump' required to formulate those premises. We identify the translation of simulation into formal axioms as the critical bottleneck in artificial scientific invention, and propose that physically consistent, multimodal world models offer the necessary sensory grounding to bridge this divide.
HN user
annapowellsmith
I'm a full-stack web developer, with a particular interest in data visualisation.
My father is in the National Survey of Health & Development study of people born in Britain in March 1946, "the longest continually running major birth cohort study in the world": https://en.wikipedia.org/wiki/National_Survey_of_Health_%26_...
He seems to enjoy it, especially as a retired research psychologist. Every few years he goes to London and they run tests, see how long he can stand on one leg, do various scans etc. They are currently running a sleep study using wearables.
The study also had a 70th birthday party in 2016 for the cohort, which I got to go to as a guest. Apparently very few people have dropped out, so hopefully it's a positive experience for most people.
There's also a book about the study: https://www.amazon.co.uk/Life-Project-Extraordinary-Story-Or...
Flourish | Senior Developer | Full-time | REMOTE (UK) | https://flourish.studio/jobs/senior-developer
Flourish is an online tool for creating beautiful data graphics, presentations and other interactive content without coding. Launched in 2018, we have grown quickly and are used by some of the world’s largest companies and news organisations.
We’re looking for a senior developer to join our team, reporting to the head of software development. You'll be developing new features for Flourish; maintaining standards for web performance, security and accessibility; building automated tests and improving the observability of the codebase; mentoring and supporting more junior developers.
Some of the technologies we use include Node.js, Express; Postgres, Redis; Handlebars, Sass; AWS; Cypress; GitHub, CircleCI, Sentry. We don’t expect you to have worked with all of these, but any you have are a plus.
More details here: https://flourish.studio/jobs/senior-developer/
Comment from a tractor-owning friend:
That's a very good piece capturing a lot of the issues (esp GPS retrofitting which is easy and make a huge difference).
The one thing it doesn't quite draw out is that the market has segmented. If you are a small to mid farmer you have fairly low demands for horsepower and can tolerate small breakdowns you can fix in the field. So buying an older machine makes sense.
If you are a big farmer or a contractor (person with a big tractor/combine/machine who hires out) you need high horsepower and 18 hour a day operation without breakdowns. as in fleet purchasing for trucks they have rolling finance for high capital value kit to replace it every couple of years having run up high hours.
The modern electronic high power kit is quite amazing, it's not gratuitous use of electronics. Electronics also important for meeting modern emissions regs.
Tractors, as the piece points out have to be ultra rugged for field work and that means that the older stuff does last well and 100hp or over will get you a long way.
Article is paywalled, so here are the key bits:
Tens of thousands of political ads have gone missing from Facebook’s public archive less than 48 hours before the British general election, raising fresh questions over measures introduced to address transparency concerns in online political advertising...
...On December 5, Facebook’s ad library, which includes adverts shown on its photo messaging subsidiary Instagram, showed a total of just under 180,000 classified as political in Britain since October 2018. But on Tuesday, that number had suddenly dropped by a third, to fewer than 120,000 ads.
Facebook declined to say what had caused the issue, or whether the ad data were irretrievably lost or could be recovered. But a company spokesperson said: “We’re aware that people are having trouble accessing the ads in the ads library, and we’re working to fix the issue as soon as possible.”...
...The data lost by Facebook did not appear to be limited to any particular party or subset — ads were missing from all the main parties’ pages.
The Professional Pilots Rumour Network:
Which is pretty much what the name suggests. Airline pilots, both passenger and commercial flights, swapping news and rumours.
London, UK | Flourish (data visualisation) | Lead Developer | ONSITE (part-time remote ok) | https://flourish.studio/jobs/lead-fullstack-developer/
Flourish is a next-generation platform for data visualisation and storytelling. It brings the power, flexibility and beauty of D3 and WebGL to anyone with a spreadsheet or database.
Flourish is being created by the award-winning visualisation studio Kiln with the backing of well-known investors in the UK and US.
We’re looking for a talented developer to join our team in London and lead full-stack JavaScript development, reporting to the CTO.
This is a key role which will involve overseeing the maintenance and development of the backend of Flourish – a web app written in Node/Express, backed by a PostgreSQL database, running on AWS.
Competitive salary, share options, 30 days’ annual holiday. Based in London, but we’re open to part-time remote work.
To apply: Send a CV and/or portfolio to jobs@kiln.digital - and email with any questions.
This re-submission has more comments, including from the developers: https://news.ycombinator.com/item?id=13164058
The Indonesian government's attempts to stop burning are hindered by a lack of proper land ownership data. From the Jakarta Post:
"The one-map policy — a comprehensive map of land ownership to provide clarity on the exact boundaries of land owned by companies — should be fully implemented. Experience over the past decades has shown there is an acute lack of transparency when it comes to maps on land ownership. Without a centralized, public map, the task of pinpointing errant companies or landowners becomes murky.
"The government also seems to have been reluctant to make its existing concession maps publicly available to public forest-monitoring platforms. If we are serious about tackling the appalling air pollution, the government must release data that will facilitate public monitoring."
http://www.thejakartapost.com/news/2015/10/27/editorial-prot...
Thanks for the suggestion! No, it's not mine - I don't know who runs it. Just thought it was a great use of data.
Really nice.