Hacker News new | past | comments | ask | show | jobs | submit | mjirv's comments login

Out of curiosity, I googled several of the pubs he mentioned. All but one* was still around.

*I found a pub called the Ranelagh, but it’s not in Pimlico, so I assume it’s a different one. It was the one he described as “really terrible,” so no big loss, I suppose.

Addendum: the other interesting thing I noticed was the ones he derided as having been “modernized” in the 1960s were also newly renovated today, with airy, Scandinavian, 2020s aesthetics. Presumably because unlike the traditional pubs, the 60s style became dated pretty quickly.


There is a pub called The Ranelagh in Bounds Green, North London which is near to what used to be a Middlesex Polytechnic site where the computer centre was located (DEC 10, two IBM 4381s, several VAXen and a couple of Primes) and where I worked in the mid to late 1980s. It was a hole then (still there, but I haven't been in for many years), but that didn't stop us programmers drinking there.


Much denser world network now


re: dbt, it looks like Hashquery is more for running analyses than doing in-data-warehouse transformations.

Unless I'm mistaken there's no native way in Hashquery right now to materialize the results back to your warehouse like you would with dbt.


Current OpenAI job postings top out at $385,000 for full stack engineers which is 98th percentile in the US (and less for San Francisco).

The €80k is the start of the Mistral salary range. €100k is the top, which would put you in the top 1% of earners in France.


The $385k is base salary, not total comp, right?


> Three out of every four desk workers report working in the 3 to 6pm timeframe, but of those, only one in four consider these hours highly productive.

It’s nice to see some data on this. A lot of people talk about a 4-day workweek, but personally I think a 6-hour work day would be better.

At least in white-collar jobs, I think most people could fit the same amount of productivity into fewer (focused) hours.

Plus a day that’s roughly 9-3 instead of 9-5 would have the benefit of being aligned with school schedules (at least here in the US) and make things easier for working parents.


> make things easier for working parents.

During the (now semi-aborted) back-to-office push at my employer, several of us pointed out that a 9-5 in the office is impossible for those of us with children and a commute.

- 8:00am - earliest free drop-off time at school.

- 8:12 or 8:20 - next bus or train to the city. I can probably make that from the school, if I fully sprint for at least half that journey. (Taking a car is riskier; traffic around the school or the station, or bad parking, means that driving is wildly more variable.)

- 9:45 - earliest possible time I can roll into the office.

- 5:00 - I slam my laptop shut, having already put on my coat, and sprint to the subway and/or bike share to head to the train station.

- 5:16 - the earliest possible train I can make, assuming that all of Midtown stops for me.

- 6:31 - the earliest possible arrival back home, 31 minutes past aftercare's closing time.

The only way this would be doable is by hiring childcare for both mornings and afternoons, every day of the week, to the tune of ~$50 extra per day, minimum. That's on top of the cost of the commute.


When I was a kid, I would get dropped off early for “before school programs at school” and stay at “after school programs” at school and get picked up late because both of my parents worked full time. Do they not have those anymore (which sound like less money than “hiring childcare”)?


Many times the slots are all taken. Then there's the cost. These things cost money. Add commute spending and many people can't really afford both commute for them and morning and afternoon clubs for their kids.


They do. But they both cost money. My child is already enrolled in the after-school program, from 2:30 until 6:00.


How much do those cost these days compared to say, day care programs (per hour)?


I'm not sure. We pay a yearly rate, and it's significantly cheaper than it would be to pay per hour.

In any case, we already pay for the aftercare regardless of whether we're commuting to work that day or not; that's a cost that's already budgeted for.


In this case it just makes sense to have RTO at reduced hours. Make the standard work day something like 10-4.

I never understood why people throw the baby out with the bathwater on things like this.


I think in our case it was driven by the company's new owners not being based in our area, and not understanding that our commute and living situations are vastly different from theirs.


Or understanding but deliberately making advantage of the difference.


well, sure if you live so far away that you spend 3 hours a day commuting... I would just work on the train and arrive late /leave early.


Sure, unless the mandated RTO plan explicitly states that you should be in the office from 9-5.

If I had to regularly spend 3 hours a day commuting, I would quit and find a job that doesn't do that shit to me.


how did you do that before covid when working from home wasn't yet common?

i understand the problem for a single parent. in most countries they just can't work 8 hours a day. the time where kids are away for me for example is barely 8 hours by the clock. that is without commuting time and lunch break. once the kids come home someone needs to start preparing dinner, followed by housework.

with two parents however, one parent could work early, and the other could work late so that there is one parent around in the morning and one in the evening.

the trouble now is of course that this becomes a coordination problem with two different employers who have different expectations for the work schedule


> how did you do that before covid when working from home wasn't yet common?

Everything was different than the scenario I described in my earlier comment, including my employer. (And to be clear, my current employer dropped the 9-5 requirement when they realized they wouldn't have any engineers left if they tried to enforce it.)

The commute was significantly shorter, and we relied on the subway - not trains - which came more regularly.

There was no requirement for a 9-5, and so I could either do both drop-off and pick-up, or we could divide the responsibilities between my spouse and I.

Furthermore, we could work from home as necessary.

> with two parents however, one parent could work early, and the other could work late so that there is one parent around in the morning and one in the evening. the trouble now is of course that this becomes a coordination problem with two different employers who have different expectations for the work schedule

Exactly. And I would have choice words for an employer who suggested that the problem they're plopping in my lap for no reason should suddenly become my spouse's concern.


To add to your example, UK school hours are (or were when I was going to them) 09:00-15:30.

I'm not 100% sure the hours here in Germany (sadly lacking the need to know), but I think many are 08:00-13:30.


Unlikely. Non-poaching agreements are illegal under federal antitrust law.


Nonsense.


The key line:

“Microsoft has assured us that there are positions for all OpenAl employees at this new subsidiary should we choose to join.”


I think everyone assumed this was an aquihire without the "aqui-" but this is the first time I've seen it explicitly stated.


hostile takeunder?


Love it. Could also be called a hostile giveover, considering the OpenAI board gifted this opportunity to Microsoft


That's perfect.


You win


will they stay though? what happens to their OAI options?


Will their OAI options be worth anything if the implosion continues?


yeah but threatening to quit is actually accelerating the implosion


I don’t believe startups can have successful exits without extraordinary leadership (which the current board can never find). The people quitting are simply jumping off a sinking ship.


MSFT RSUs actually have value as opposed to OpenAI’s Profit Participation Units (PPU).

https://www.levels.fyi/blog/openai-compensation.html

https://images.openai.com/blob/142770fb-3df2-45d9-9ee3-7aa06...


What will happen to their newly granted msft shares? One can be sold _today_ and might be worth a lot more soon…


Sounds a lot like MS wants to have OpenAI but without a boards that considers pesky things like morals.


Time for a counter-counter-coup that ends up with Microsoft under the Linux Foundation after RMS reveals he is Satoshi...


You mean the GNU Linux Foundation?


RMS (I assume Richard Stallman) may be many many many things, but setting up a global pyramid scheme doesn't seem to be his M.O.

But stranger things have happened. One day I may be very very VERY surprised.


There is nothing related to pyramids in bitcoin. It's just an implementation of a novel, trustless electronic money, also it's free software.


how would you define an asset that has zero intrinsic value other than the value people have already committed to it? house of cards?


Money.


but money has intrinsic value. It is directly tied to the economy of a country. So if a country was to collapse, so would it's money.

The point here is that if a country collapses, then you got bigger problems than the loss of whatever stored currency you got. Even if your money is in the hypothetically useful crypto, you got far bigger problems that the money you own is useless to you, you need to survive.

But aside from that extreme scenario, money is not the same thing.

Another way to think of it:

There is nothing in the world that would prevent the immediate collapse of crypto if everyone who owns it just decided to sell.

If everyone in the world stops accepting the US Dollar, the US can still continue to use it internally and manufacture goods and such. It'll just be a collapse of trade, but then even in that scenario people can just exchange the dollar locally for say gold, and trade gold on the global market. So the dollar has physical and usable backing. Meanwhile crypto has literally nothing.


There were many currencies in history that have lost all or almost all its value upon serious economical crisis in respective countries. It seems you wouldn't call that money? Crypto is simply an alternative currency.

See also: https://en.m.wikipedia.org/wiki/Private_currency


Right, INTERNAL economical crisis is what causes the collapse of currency. But just because the rest of the world doesn't recognize it, doesn't mean it is worthless, it simply converts.

Bitcoin has nothing in and of itself.

Also private currency like script was awful, please don't take the worst financial examples in history and claim that bitcoin is similar as an argument as to why it is valid.


> Bitcoin has nothing in and of itself.

I don't understand what is "in and of itself" in an ordinary currency of an ordinary, small country.

> INTERNAL economical crisis is what causes the collapse of currency

Which is why it is very unlikely to happen with bitcoin.

> But just because the rest of the world doesn't recognize it, doesn't mean it is worthless, it simply converts.

Can't you say exactly the same about bitcoin?


The year of the Linux Microsoft.


again, nobody has shown even a glimmer of the board operating with morality being their focus. we just don't know. we do know that a vast majority of the company don't trust the board though.


Sam just gave 3 hearts to Ilya as well... I hope the drama continues and he joins MS at this point.


Whose morals again?


That is a spectacular power move: extending 700 job offers, many of which would be close to $1 million per year compensation.


They didn’t say anything about the compensation.


So essentially, OpenAI is a sinking ship as long as the board members go ahead with their new CEO and Sam, Greg are not returning.

Microsoft can absorb all the employees and switch them into the new AI subsidiary which basically is an acqui-hire without buying out everyone else's shares and making a new DeepMind / OpenAI research division inside of the company.

So all along it was a long winded side-step into having a new AI division without all the regulatory headaches of a formal acquisition.


> OpenAI is a sinking ship as long as the board members go ahead with their new CEO and Sam, Greg are not returning

Far from certain. One, they still control a lot of money and cloud credits. Two, they can credibly threaten to license to a competitor or even open source everything, thereby destroying the unique value of the work.

> without all the regulatory headaches of a formal acquisition

This, too, is far from certain.


>Far from certain. One, they still control a lot of money and cloud credits.

This too is far from certain. The funding and credits was at best tied to milestones, and at worst, the investment contract is already broken and msft can walk.

I suspect they would not actually do the latter and the ip is tied to continual partnership.


And sue for the assets of OpenAI on account of the damage the board did to their stock... and end up with all of the IP.


On what basis would one entity be held responsible for another entity’s stock price, without evidence of fraud? Especially a non profit.


The value of OpenAI's own assets in the for-profit subsidiary, may drop in value due to recent events.

Microsoft is a substantial shareholder (49%) in that for-profit subsidiary, so the value of Microsoft's asset has presumably reduced due to OpenAI's board decisions.

OpenAI's board decisions which resulted in these events appear to have been improperly conducted: Two of the board's members weren't aware of its deliberations, or the outcome until the last minute, notably the chair of the board. A board's decisions have legal weight because they are collective. It's allowed to patch them up after if the board agrees, for people to take breaks, etc. But if some directors intentionally excluded other directors from such a major decision (and formal deliberations), affecting the value and future of the company, that leaves the board's decision open to legal challenges.

Hypothetically Microsoft could sue and offer to settle. Then OpenAI might not have enough funds if it would lose, so might have sell shares in the for-profit subsidiary, or transfer them. Microsoft only needs about 2% more to become majority shareholder of the for-profit subsidiary, which runs ChatGPT sevices.


Bad Faith. Watch the sales presentation that Altman and Nadella gave at OpenAI’s inaugural developer conference just a few days/hours before OpenAI fired its key executives, including Altman.


No. As one of the other commenters mentioned, Sam (and possibly Greg) probably recused himself and didn't vote (likely forced to by the board's bylaws).


So maybe it was a 3-1 vote with Ilya voting against? That would be infuriating.


infuriating, but man it's gonna make a great HBO series


Nice! We could be in the market for something like this soon.

What’s your quick pitch for why someone should use Revert vs competitors like Merge.dev and Nango?


* vs Nango - we think they have great support for OAuth management and do support more APIs than ours on the surface (atleast at the moment).

Nango is based on Pizzly an existing OSS project which they built on top of. We're building it from the ground up.

Even though they seem to have more integrations, our integration support is better than them in terms of the depth of use-cases allowed (more standard objects supported, custom properties, field mapping support, custom objects (soon) etc).

A few prospects of ours tried out Nango for this use-case and then came to us eventually.

* vs Merge - we'd be able to fly past the number of integrations offered by them being an OSS product especially because of community contributed integrations. Being a developer first product, open-source is the way to build the best product in this category.

Integrations inevitably have edge cases that you would run into and you as a customer might require Merge to behave in a certain way. The typical response at a closed-source SaaS company would be that its on "their roadmap", never to get back again. This holds you tightly with their roadmap velocity and you're locked into a vendor.

Being an open-source product you, the engineer, will be able to fix or add integrations right away in the worst case if nothing else. This way of operating is very powerful we think.

Also, we don't cost you an arm and a leg :)


I love these kinds of products, and welcome any competition in the space. But, this comparison to Nango doesn't seem accurate, so I feel inclined to comment.

Please correct me if I'm wrong, but you say...

> Even though [Nango] seem to have more integrations

Nango has north of 100 integrations, Revert seems to have 4 atm?

> our integration support is better than them in terms of the depth of use-cases allowed (more standard objects supported, custom properties, field mapping support, custom objects (soon) etc).

How so?

Nango Sync gets you easy access to the raw API responses from the 3rd party service, and lets you map that to whatever shape/model you, as the implementer, want to end up with.

Revert seems to return standardized/normalized objects per data model (e.g, company, contact, task) across the 4 different integrations currently mentioned. It also seems to support "custom mapping" past the "lowest common denominator" schema, by adding `sourceFieldName` -> `targetFieldName` mappings (but seemingly only for picking out response key if they're strings, not any "pick from object", or "compute based on multiple properties"?)

Please don't take this as discouragement -- it's a great space to play in, and there's a lot of room for improvement. But, as a _very_ happy user of Nango over the past 10+ months, I feel you should compare yourself honestly at the very least.

Good luck!


Hi! Thanks for your comment.

> Even though [Nango] seem to have more integrations

We agree Nango has more integrations and we love OSS software so I'm with you on this. Credit where credit is due and we don't want to make false claims at all. We never claimed to have more integrations than them. I'm not sure how what I posted came off as dishonest.

> but seemingly only for picking out response key if they're strings, not any "pick from object", or "compute based on multiple properties"?)

I'd say we support this perhaps in a different way.

I have not used Nango myself to comment on specific ways it handles data vs how we handle it.

Its great that you're liking Nango and we want OSS/better product to win regardless.


Yeah, sorry, I just got caught up in your wording. Since you asked: "Nango seems to have more integrations" feels disingenuous, when you're comparing 4 to 100+. You'll likely be asked to compare yourself with Nango a lot, so it's not a bad idea to know what you're up against.

In any case, I wish you the best of luck with the "one model per resource type" concept you're trying. It's a tricky one, since you're usually stuck with the lower common denominator.

I expect many, if not most users will need additional custom mapping (so if "field A" -> "field B" mapping is the only option for now, expect to run into lots of feature requests that need to pick from objects/compute multiple values into one field. DX around this will be important)


https://www.nango.dev founder here.

I think the biggest difference is that Nango lets you customize & extend the unified APIs on the platform.

Usually unified APIs mitigate their limited catalog with passthrough/proxy requests. But this is a partial solution, since you go back to having a lot of integration logic in your code base.

With Nango these customizations live in the unified API itself and benefit from all the infrastructure available there (OAuth, rate-limit handling, pagination, de-duplication of records, etc.). You can also build entirely custom integrations in Nango.

That being said, I think open-source unified APIs have a ton of promise! It is great to see the ecosystem grow :)


We agree on the promise for sure! :)


Interesting that they don't have the watermark on the "DALL-E 3" images.


Perhaps they use some form of steganography.


Why would you expect one? I don't think I've heard of any AI image generators placing a watermark.


v2 has a watermark in the lower right corner of every image.


Thats the case now too if you use the API and not the labs version


why do you think this is?


There's probably no point in terms of a watermark's goals.

Additionally, since OpenAI is asserting no rights over the image, a watermark would be inappropriate.


Makes sense


Nice. We (Delphi[0] - enterprise AI data assistant) do something similar internally to power our app.

We’re all typescript under the hood so I’ll give this a look and see if we can use it.

Symphony wouldn’t support other LLMs currently, right? Only GPT-4?

[0] https://delphihq.com


Glad to hear that!

Right, currently Symphony only supports GPT-4 and GPT-3.5-turbo since they're the only ones with native API support.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: