chrstphrhrt

Haha, oh well I'd still try one with more filling, it is my favourite part anyway. Canada Post is pulling some culinary activism I can get behind. Thanks for looking past the rockies Ottawa!

Down in Canada (Quebec).

We should fund medieval literature well too. Some of the best CTOs are humanities people for whom engineering just isn't fully satisfying at intellectual life scale.

Applying symbolic logic, hermeneutics, retroduction, and criticism are higher abstract levels of creative problem solving, via inference to the likely explanation with respect to priors, a.k.a. the scientific method. These kinds of reasoning simply aren't practically doable by math alone yet, not to mention deep learning tools in the field.

The next wave of AI methods will need to go beyond statistics to come full circle, because classification is not reasoning. What else is there? All the GOFAI stuff, prolog, expert systems, knowledge graphs, fuzzy/modal logics. These things were always great at reasoning but the last winter happened because they were bad at dealing with uncertainty. Now the stats methods have solved that problem but they suck at reasoning. It is time for a synthesis.

As a startup I don't want a data science PhD to tinker with an optimal CNN design for a year to eke out that extra 1 or 2 percent of precision and recall. I want an ensemble of specialized models, ideally pre-trained, grafted together in a logically consistent way that is configurable and introspect-able via text/data at runtime.

The economy is misconfigured! That said, humanities people should probably learn how to code as a research skill.

Could it be https://en.wikipedia.org/wiki/Henry_George ?

Also curious about techniques for re-translating enhancement proposals back into business terms and updating high-level requirements based on capabilities discovered in the research/prototyping for the designs, if anyone can spare some wisdom on that side.

Would keeping some kind of technical functionality-to-sales features/benefits matrix be a worthwhile exercise?

Similar profile, would not considering working for them. My first and only interview with FB a few years ago I nailed the initial hoop jumping about IPC and other *nix things but overall just got a really bad taste in my mouth. Also went for dinner, which was nice but still, something is creepy about them. They are looking for some kind of desperation that I just don’t have. Sounds like some fun problem solving though.

Haven’t had a chance to hire any ex-FAANGers yet, do they compare well to the general market?

We need to bring back abductive/retroductive inference (to the best explanation) as if it were its own mode of reasoning to complement deduction and induction. This is how scientific discoveries are often made IMHO but often chalked up to some kind of intuition. There is actually a logical structure even though is is hard to formalize and prove.

https://en.wikipedia.org/wiki/Abductive_reasoning

I am working on dialog management stuff as an engineer who sits between data scientists, NLP, and product people. It is understandable that trained stats people see everything as a stats problem (induction), but we forgot about GOFAI & symbolic AI (deductive) systems along the way.

Philosophy has actually been ahead of the curve on this one for over 140 years. Maybe the "3rd wave" of AI will be to synthesize.

https://www.darpa.mil/attachments/AIFull.pdf

I personally have not had a chance to read Norvig or dive into LISP or Prolog much, but am using CLIPS expert system in a project and have toyed with core.logic in Clojure back in the day. Lots of opportunity to resurrect some of these older methods that may be superior to imperative/bespoke code given modern hardware and infrastructure.

Actually left due to anti-engineering culture (no tech leaders), succumbed to the distortions of Conway's law due to imbalance between professional services and core/platform.

That said, the nature of the lack of leadership was around the MBA mindset that data science is the be-all end-all of AI, which IME is so far from the truth.

There needs to be many layers/spirals of naive methods (i.e. straight ahead engineering) as a vanguard on the front lines with the DS/NNs/NLP bringing up the flanks for aspects of the domains that become more well understood over time.

So yeah, tail wagging the dog, cart before horse etc. If you just barrel ahead with a monolithic NN pretending that all classes are created equal (structureless blob), which is what our DS PhDs were doing, it will quickly ossify. In our case it was intent prediction so we got to pretty high accuracy, but the labels were indicating many different categories that actually had relations in nature that could not be expressed. Ironically it took some regular engineers to research HMC (hierarchical multi-label classification) and ensembles and implement a new model training framework to support them. Not sure what they teach in school but it doesn't seem to be very practical.

EDIT: now working on messaging (SMS/MMS/RCS/ABC) with a focus on dialog management and logical rules with layering-in of progressively less naive stats methods strategically rather than blindly. It helps to have an existing revenue stream that can be leveraged to create even more value with AI features rather than the backwards hail Mary of not iterating from a foothold of existing traction.

Can confirm, worked at an AI hype company for over a year and built all their systems outside the NN internals. There is going to be a correction.

Ditto.

Also when someone is barely getting by, assuming they are not dealing with addiction or violence and generally just subsisting, you can bet they know exactly what to do with that last $20. E.g. groceries with the highest calories per dollar, how many bus passes to get at a time, etc.

If budgeting is what is meant by investment (since obviously we're not talking about saving if there's nothing to be saved), from what I have seen poor people are incredibly good at it.

Contrast that with money managers who are generally just bad and can't outperform the market, not for lack of trying. Is "investment" really the right idea, and who says it's the number one importance compared to basic human needs?

Why does everything have to be about abstract efficiency and cost? It's not like there's a shortage of business opportunities for techno-luxuries that are fun to produce and trade, that we absolutely must go after basic rights via austerity. Lousy rentier mindset, we should have to be creative to profit. Once monopolies emerge they should be broken up, and once technologies become necessary for daily life, access to them should become a right. Tax incentives to invest in R&D instead of hoarding wealth could be expanded. So many things can be done to mitigate issues within "social democracy" without going full public ownership or eliminating private property. Question is, once these things are implemented, how can their repeal be prevented? Time to open up and revise constitutions.

Awesome! Sorry out of the loop since only use them when visiting the states.

Vancouver is so backwards :(

As a customer experience I prefer Lyft because they let me give tips. I doin't think Uber drivers make a living wage, so I like trying to figure out what can be added to offset the low base/fees so the drivers can do okay.

Can only do this if able to build up a state of flow around it. Start small, as small as possible. Not everything good has to be big.

All the comments like "dump everything and dive in head first for X months" sound like some weird cult of extreme heroism to me.

Everything is a process, not some abstract form that can magically be reified into existence by wanting/trying harder. We are animals and have to build up the habit and curiosity in tandem, else risk burnout.

She's amazing. So humble and accessible.

I agree. Especially if there are other teammates who defend the bad codebase because of their association to it.

Nobody should claim washing dishes is inherently wasteful or pointless. Legacy code on the other hand, comes with a lot of social pressure to just layer on more debt.

I wasn't, thanks will get that patched up!

Quite informative, thanks.

There goes my afternoon :)

I read at some point that LVB was a deist and a copy of the Enneads of Plotinus was found in his library after he died.

Just listened to some of this Reicha dude's stuff on Spotify and it is so similar to Beethoven thematically.

I wonder if they were both going for the same themes of ideal oneness vs. multiplicity, if they had shared approaches to exploring the space of musical theory in a neoplatonic way. Given LVB's love of wine I can only imagine they got pretty nerdy and unhinged.

Love LTREE! Used it for domain label hierarchies (multiple roots - an "ontology") for a data science team to be able to train many different models with shared/overlapping label sets (an ensemble of models for various domains). It makes queries to get training data powerful since you can ask for "all descendants of X" or "ancestors of Y where Z".

It's really nice!

However, I can't stop thinking "tuberculosis".

Is that why smoking weed makes things seem profound, from instantly forgetting the less salient stuff?

While I agree with the fact that the Enlightenment helped to get rid of some harmful superstitions, I think it also contributed to the mechanistic industrial mindset that ended up causing a lot of harm e.g. colonialism and eugenics.

Great list though, maybe add on reproductive justice and climate science.

Awesome! Memories of "hacking" that gorillas-throwing-bananas game to mess with the physics, probably around 1995.

Not just you. It's a thing.

Even more extreme, consider that companies that do charge money drop the ball constantly on security/privacy, and basic quality. And yet they continue growing, maybe having a bit of churn.

I think there's something about generosity that just attracts entitled bullies/leeches because they see it as an opportunity to "win" against people who are "foolish" (aka "weak") enough to work for free. Sadly sometimes the generous are duped into considering such requests out of curiosity or politeness, which leads to burnout or worse.

Ever notice how bullies will always memorize these small personal details about a target and then when manipulating or gaslighting make sure to mention those facts?

E.g. Oh a new crisis "just came up". I need you to complete this big menial data entry task that I forgot about for a week right now (drop whatever you were working on before). Don't you want to make it to your daughter's concert tonight? Best get it done by EOD so you can be there for her. Next day: how's progress on that thing you were working on before?

Stored procedures! Sounds fun but does anyone know a practical use that can't be achieved with normal PL/pgSQL or other embedded scripting languages, plus regular functions?

This should include compensation ranges.

I really enjoyed this writeup. Having just finished a year in the field on the engineering side, and reading many papers without an academic background, it’s funny to recognize who are the “math” types and who are the “biology” types at work. Puts things in perspective.

I’ve been pushing for more ensembles and multi-label classifiers because I want to orchestrate the pieces with logic to fill gaps until stats methods outperform (given enough new data in nice annotation-friendly structures).

The “math” types seem to feel like there’s a neural net solution to every problem or that we can expand the multi class model to cover more domains despite being mostly saturated with high accuracy. Sounds awesome but I’m impatient!

The “biology” folks seem to be most attracted to neatness or parsimony. We had some great bikeshedding sessions around adjacency list vs materialized path (ltree) in postgres for label hierarchies. Abstractions can be useful too!

Any tips on being a better experimentalist and pushing academic colleagues towards better solutions in the field?