mbb70

u7à

YMMV but every gym I've been to in the US had a huge social component.

If the unit-economics work out and they can sell $0.99 of tokens for $1.00, doesn't matter how many agents you spin up. The flat rate subscriptions can't last though.

Reminds me of "a terrible project with a great slide deck might end up decent. A great project with a terrible slide deck won't even exist."

In the real world there is no If You Build It They Will Come, you've got to get the word out

https://news.ycombinator.com/newsguidelines.html#generated

They are a luxury item, you are paying for the privilege of signaling you can afford $550 headphones. Generic black over-ear headphones could be $800, could be $80, useless for signaling. Doubly true in the context of a gift.

https://news.ycombinator.com/newsguidelines.html#generated

> "almost no competition"

> "costs way too much"

> "It's not like chat is a hard problem"

Surely these statements can't all be true. Since Slack is expensive and has little competition, I think chat is a harder problem than you think.

Ads are a ratchet that only tighten in one direction. Once the paychecks of 1000s of motivated, intelligent OpenAI employees depend on ad revenue increasing, the only option is to make them more invasive, more prevalent, more annoying, more data hungry etc.

In the foreward, she calls out to her, great SF is descriptive, not predictive. TLHOD is about sex, gender, friendships and culture in our world.

Also a huge number of spacey contemporary works like A Mote in God's Eye, Rendezvous with Rama, Dune, Ringworld...

That's not fair! Sometimes the ideas come from Snow Crash, which gave us the Metaverse because Zuckerberg wanted to cut a guy in half with a katana from a motorcycle.

Enshittification is a technique to make more money _without_ raising the price by simply making the product worse.

Self-driving taxis have a high floor for 'making the product worse' because the car fundamentally has to drive itself.

As is often stated, microservices is a solution for scaling an engineering org to 100s of developers, not for scaling a product to millions of users.

Start with your chosen dictionary, remove small words and all 's'es.

Then find all words with 7 unique letters, then get the unique set of letter sets. These are your valid puzzles times 7 for each selection of a center letter.

Construct a trie of your dictionary. For each letter of each puzzle, walk the trie using only the puzzle letters. When you find a word, if it used the center letter, add to list.

In practice it's log(n) for size of dictionary.

The bigness of your data has always depended on the what you are doing with it.

Consider the following table of medical surgeries: date,physician_name, surgery_name,success.

"What are the top 10 most common surgeries?" - easy in bash

"Who are the top physicians (% success) in the last year for those surgeries?" - still easy in bash

"Which surgeries are most affected by physician experience?" - very hard in bash, requires calculating for every surgery how many times that physician had performed that surgery on that day, then compare low and high experience outcomes.

A researcher might see a smooth continuum of increasingly complex questions, but there are huge jumps in computational complexity. At 50gb dataset might be 'bigger' than a 2tb one if you are asking tough questions.

It's easier for a business to say "we use Spark for data processing", than "we build bespoke processing engines on a case by case basis".

Probably it actually ran git checkout or reset. As you say git revert only operates on committed snapshots so it will all be in the reflog

I explored the space of valid Spelling Bee puzzles and found out the lowest scoring puzzle is (x)bejkou with 14 points.

Hoping they do it for April 1st one year.

Wish we lived in the universe where the term 'monkey' won over 'agent'. Would have given everything a cool Planet of the Apes feel.

I remember this getting a lot of buzz at the time, but few orgs are at the level of sophistication to implement chaos testing effectively.

Companies all want a robust DR strategy, but most outages are self-inflicted and time spent on DR would be better spent improving DX, testing, deployment and rollback.

The goal of "actual projects" is also fast and correct enough though

The very next ask will be "order the zipcodes by number of customers" at which point you'll be back to aggregations, which is where you should have started

"Figure out which codes they can use to get the most revenue" is a billion dollar industry with many players, subspecialties and surprisingly few lawsuits.

'robot' meaning 'automated/non-human', a common English idiom especially in non-technical contexts

Really feels like we owe KnockoutJS an apology

Long wait times are not exclusive to public healthcare. My dermatology appointment to examine a concerning mole was scheduled 9 months out. And of course I pay for the privilege.

I implemented a casino in assembly for college. Started with a Mersenne Twister and added a few pure chance games like roulette and slots.

The PRNG was trivial. Managing the user's bank balance, switching between games, making the roulette wheel 'spin' by printing a loop of numbers slower and slower was painstaking, error ladden work.

State doesn't just contributing to complexity, it is complexity.

A huge and growing percent of all realtime communication is happening via text.

It is reasonable and worthwhile to encode some nonverbal information in it, and emojis have won the day.

The key is goodness/badness of advice is a function of the receiver. The internet doesn't give you control over who reads your stuff, so internet advice is safer and less useful than it could be.

The advice "use 'any' if it's too much work to type" is dangerous/bad advice for some developers because they don't have a well tuned definition of 'too much work', and they might not have all the tricks in the toolbox for every situation.

But legacy code or poorly typed libs can be an infinity time suck, and the most pragmatic approach might be to cut your losses, slap an 'any' on it and move on.

A great mentor gives the best (different axis than good/bad or safe/danger) advice for an individual in a specific situation.

In healthcare, notes are directly correlated with $$$ for the hospital, because everything that is billed for must be documented with a mix of metrics (O2%, temp, lab results), events (orders, prescriptions, procedures) and notes (consultation notes, imaging interpretations, discharge summaries).

Billions get spent annually in administrative overhead focused on squeezing the most money out of these notes as possible. A tremendous expense can be justified to increase note quality (aka revenue, though 'accuracy/efficiency' is the trojan horse used to slip by regulators).

GenAI has a ton of potential there. Likewise on the insurance side, which has to wade through these notes and produce a very labor intensive paper trail of their own.

Eventually the AIs will just sling em-dashes at each other while we sit by pool.

The status page is a SaaS product called StatusPage, acquired by Atlassian in 2016.

Normally just called calendar versioning