News for hackers. Stuff that matters.
Re:OpenAI's employees were given two explanations for why Sa... by kumarvvr on Monday November 20, 2023 @08:42PM (#38357711)
Giving 2 people the same project? Isnt this like the thing to do to get differing approaches and then release the amalgamation of the two? I thought these sorts of things are common.

Giving different opinions on same person is a reason to fire a CEO?

This board has no reason to fire, or does not want to give the actual reason to fire Sam. They messed up.

Re:OpenAI's employees were given two explanations for why Sa... by stingraycharles on Monday November 20, 2023 @09:13PM (#38358040)
I remember a few years ago when there was some research group that was able to take a picture of a black hole. It involved lots of complicated interpretation of data.

As an extra sanity check, they had two teams working in isolation interpreting this data and constructing the image. If the end result was more or less the same, it’s a good check that it was correct.

So yes, it’s absolutely a valid strategy.

Re:OpenAI's employees were given two explanations for why Sa... by msravi on Monday November 20, 2023 @10:25PM (#38358700)
Did the teams know that there was another team working on the same thing? I wonder how that affects working of both teams... On the other hand, not telling the teams would erode the trust that the teams have in management.
Re:OpenAI's employees were given two explanations for why Sa... by Keyframe on Tuesday November 21, 2023 @12:48AM (#38359867)
There were four teams actually. They knew but couldn't talk to each other. There's a documentary about it. I highly suggest watching it, it also features late Stephen Hawking et al. working on black hole soft hair. Documentary is called Black Holes: The Edge of All We Know, it's on pretty much all streaming platforms.
Re:OpenAI's employees were given two explanations for why Sa... by campbel on Monday November 20, 2023 @09:46PM (#38358347)
Yep! I've done eng "bake-offs" as well, where a few folks / teams work on a problem in isolation then we compare and contrast after. Good fun!
Re:OpenAI's employees were given two explanations for why Sa... by cyrnel on Monday November 20, 2023 @11:44PM (#38359370)
Not good fun when it's done in secret. That happened to me, and I was gaslit when I discovered the competing git repo.

Not saying that's what happened here, but too many people are defending this horrid concept of secretly making half your workers do a bunch of work only to see the boulder roll right back down the hill.

Re:OpenAI's employees were given two explanations for why Sa... by moorow on Tuesday November 21, 2023 @12:00AM (#38359493)
Yeah, doing it in secrecy is a recipe for Bad Things. I worked at a startup that completely died because of it.
Re:OpenAI's employees were given two explanations for why Sa... by campbel on Wednesday November 22, 2023 @12:36AM (#38375102)
Yeah that sounds toxic, this was done with everyones knowledge
Re:OpenAI's employees were given two explanations for why Sa... by DonHopkins on Monday November 20, 2023 @10:46PM (#38358876)
Maybe they needed two teams to independently try to decode an old tape of random numbers from a radio space telescope that turned out to be an extraterrestrial transmission, like a neutrino signal from the Canis Minor constellation or something. Happens all the time.

https://en.wikipedia.org/wiki/His_Master%27s_Voice_(novel)

Re:OpenAI's employees were given two explanations for why Sa... by hal009 on Monday November 20, 2023 @10:16PM (#38358623)
As mentioned by another person in this thread [0], it is likely that it was Ilya's work that was getting replicated by another "secret" team, and the "different opinions on the same person" was Sam's opinions of Ilya. Perhaps Sam saw him as an unstable element and a single point of failure in the company, and wanted to make sure that OpenAI would be able to continue without Ilya?

[0] https://news.ycombinator.com/reply?id=38357843

Re:OpenAI's employees were given two explanations for why Sa... by JimDabell on Monday November 20, 2023 @11:03PM (#38359014)
Since a lot of the board’s responsibilities are tied to capabilities of the platform, it’s possible that Altman asked for Ilya to determine the capabilities, didn’t like the answer, then got somebody else to give the “right” answer, which he presented to the board. A simple dual-track project shouldn’t be a problem, but this kind of thing would be seen as dishonesty by the board.
Re:OpenAI's employees were given two explanations for why Sa... by hn_throwaway_99 on Tuesday November 21, 2023 @01:30AM (#38360120)
> it’s possible that Altman asked for Ilya to determine the capabilities, didn’t like the answer, then got somebody else to give the “right” answer, which he presented to the board.

This makes no sense given that Ilya is on the board.

Re:OpenAI's employees were given two explanations for why Sa... by JimDabell on Tuesday November 21, 2023 @01:56AM (#38360294)
No, it just means that in that scenario Sam would think he could convince the rest of the board that Ilya was wrong because he could find somebody else to give him a preferable answer.

It’s just speculation, anyway. There isn’t really anything I’ve heard that isn’t contradicted by the evidence, so it’s likely at least one thing “known” by the public isn’t actually true.

Re:OpenAI's employees were given two explanations for why Sa... by kmlevitt on Monday November 20, 2023 @10:46PM (#38358874)
Firing Sam as a way of sticking up for Ilya would make more sense if Ilya wasn’t currently in support of Sam getting his job back.
Re:OpenAI's employees were given two explanations for why Sa... by onimishra on Tuesday November 21, 2023 @03:09AM (#38360801)
I’m not sure Ilya was anticipating this to more or less break OpenAI as a company. Ilya is all about the work they do, and might not have anticipated that this would turn the entire company against him and the rest of the board. And so, he is in support of Sam coming back, if that means that they can get back to the work at hand.
Re:OpenAI's employees were given two explanations for why Sa... by kmlevitt on Tuesday November 21, 2023 @04:17AM (#38361291)
Perhaps. But if the board is really so responsive to Ilya's concerns, why have they not reversed the decision so that Ilya can get his wish?
Re:OpenAI's employees were given two explanations for why Sa... by 015a on Monday November 20, 2023 @10:50PM (#38358911)
This is an interesting theory when combined with this tweet from Google DeepMind's team lead of Scalable Alignment [1].

[1] https://twitter.com/geoffreyirving/status/172675427761849141...

The "Sam is actually a psychopath that has managed to swindle his way into everyone liking him, and Ilya has grave ethical concerns about that kind of person leading a company seeking AGI, but he can't out him publicly because so many people are hypnotized by him" theory is definitely a new, interesting one; there has been literally no moment in the past three days where I could have predicted the next turn this would take.

Re:OpenAI's employees were given two explanations for why Sa... by doktrin on Monday November 20, 2023 @11:12PM (#38359103)
Interesting. I’m glad he shared his perspective despite the ambiguity.
Re:OpenAI's employees were given two explanations for why Sa... by nvm0n2 on Tuesday November 21, 2023 @09:46AM (#38364301)
That guy is another AI doomer though, and those people all seem to be quite slippery themselves. Supposedly Sam lied to him about other people, but there's no further detail provided and nobody seems willing to get concrete about any time Altman has been specifically dishonest. When the doomer board made similar allegations it seemed serious for a day, and then evaporated.

Meanwhile the Google AI folks have a long track record of making very misleading statements in public. I remember before Altman came along and made their models available to all, Google was fond of responding to any OpenAI blog post by claiming they had the same tech but way better, they just weren't releasing it because it was so amazing it just wasn't "safe" enough to do so yet. Then ChatGPT called their bluff and we discovered that in reality they were way behind and apparently unable to catch up, also, there were no actual safety problems and it was fine to let everyone use even relatively unconditioned models.

So this Geoffrey guy might be right but if Altman was really such a systematic liar, why would his employees be so loyal? And why is it only AI doomers who make this allegation? Maybe Altman "lied" to them by claiming key people were just as doomerist as those guys, and when they found out it wasn't true they wailed?

Re:OpenAI's employees were given two explanations for why Sa... by BillyTheKing on Monday November 20, 2023 @11:06PM (#38359044)
either that or Sam didn't tell Adam D'Angelo that they were launching a competing product in exactly the same space that poe.ai had launched one. For some context, poe had launched something similar to those custom GPTs with creator revenue sharing etc. just 4 weeks prior to dev-day
Re:OpenAI's employees were given two explanations for why Sa... by m3kw9 on Tuesday November 21, 2023 @12:16PM (#38366647)
Not sure how he would see that coming? It was a UI tweak away for OpenAI
Re:OpenAI's employees were given two explanations for why Sa... by danbmil on Monday November 20, 2023 @10:08PM (#38358537)
The CEO's I've worked for have mostly been mini-DonaldT's, almost pathologically allergic to truth, logic, or consistency. Altman seems way over on the normal scale for CEO of a multi-billion dollar company. I'm sure he can knock two eggs together to make an omelette, but these piddling excuses for firing him don't pass the smell test.

I get the feeling Ilya might be a bit naive about how people work, and may have been taken advantage of (by for example spinning this as a safety issue when it's just a good old fashioned power struggle)

Re:OpenAI's employees were given two explanations for why Sa... by danbmil on Monday November 20, 2023 @10:11PM (#38358572)
as for multiple teams with overlapping goals -- are you kidding me? That's a 100% legit and popular tactic. Once CEO I worked with relished this approach and called it a "Steel-cage death match"!
Re:OpenAI's employees were given two explanations for why Sa... by aravindgp on Monday November 20, 2023 @10:22PM (#38358683)
You were right Ilya was naive , he regrets his decision on twitter. And he was taken advantage of by power hungry people behind.
Re:OpenAI's employees were given two explanations for why Sa... by WalterBright on Monday November 20, 2023 @09:40PM (#38358295)
Back in the late 80s, Lotus faced a crisis with their spreadsheet, Lotus 1-2-3. Should they:

1. stick with DOS

2. go with OS/2

3. go with Windows

Lotus chose (2). But the market went with (3), and Lotus was destroyed by Excel. Lotus was a wealthy company at the time. I would have created three groups, and done all three options.

Re:OpenAI's employees were given two explanations for why Sa... by quickthrower2 on Monday November 20, 2023 @09:42PM (#38358319)
Which would have had been a tradeoff too. More time to market, fewer people on each project, slowed down by cross platform code.
Re:OpenAI's employees were given two explanations for why Sa... by kumarvvr on Monday November 20, 2023 @09:52PM (#38358406)
At the time, Lotus was a good company in great shape. The management could have hired people to get stuff done. In hindsight, sure, we can be judgmental, but it is still a failure in my view.

For a company selling licenses for installations, wouldn't having support for all available and upcoming platforms a good thing? Especially when the distribution costs are essentially 0?

Re:OpenAI's employees were given two explanations for why Sa... by WalterBright on Monday November 20, 2023 @11:36PM (#38359306)
Lotus was a rich company, and could have easily funded 3 full strength independent dev teams. It would not have slowed anything down.
Re:OpenAI's employees were given two explanations for why Sa... by jeremyjh on Monday November 20, 2023 @09:51PM (#38358394)
They would have just forked the code and maybe merged some changes back and forth, no real need for cross-platform code.
Re:OpenAI's employees were given two explanations for why Sa... by primax on Monday November 20, 2023 @10:01PM (#38358478)
This was pre-1983. Forking wasn't a thing at the time. Any kind of code management was cutting edge, and cross-platform shared code wasn't even dreamed of yet.
Re:OpenAI's employees were given two explanations for why Sa... by bawolff on Monday November 20, 2023 @10:19PM (#38358647)
Forking and merging is a social phenomenom. Sure git makes it easier, but nothing stopping anyone from just copying and pasting as appropriate. Not to mention diff(1) was invented in 1974, and diff3(1) in 1979, so there were already tools to help with this, even if not as well developed as modern tools.

I'm also pretty sure cross-platform code was a thing in 1983. Maybe not to the same extent and ease as now, but still a thing.

Re:OpenAI's employees were given two explanations for why Sa... by WalterBright on Monday November 20, 2023 @11:38PM (#38359323)
Successful 8086 projects were usually written in assembler - no way to get the speed and size down otherwise. I'm pretty sure Lotus 123 was all in assembler.
Re:OpenAI's employees were given two explanations for why Sa... by bawolff on Tuesday November 21, 2023 @12:19AM (#38359653)
I'm not an assembly programmer and not very familiar with how that world works, but even then, if the two OSs were for the same architecture (x86), couldn't you still have a cross OS main part and then specific parts that deal with operating system things? I normally think of compiled languages like c being an abstraction over cpu architecture, not operating system api.
Re:OpenAI's employees were given two explanations for why Sa... by WalterBright on Tuesday November 21, 2023 @12:46AM (#38359847)
Yes, you can have common assembler code among platforms, provided they use the same CPU.

From what I've seen of code developed in the 80s, however, asm code was not written to be divided into general and os specific parts. Writing cross-platform code is a skill that gets learned over time, usually the hard way.

Re:OpenAI's employees were given two explanations for why Sa... by jeremyjh on Monday November 20, 2023 @10:40PM (#38358818)
You make a copy of the files and work on them and that is a fork.
Re:OpenAI's employees were given two explanations for why Sa... by adastra22 on Monday November 20, 2023 @11:06PM (#38359043)
How do you merge changes between the source trees?

Keep in mind this predates basically ANY kind of source control. It would have been nearly 3x the work.

Re:OpenAI's employees were given two explanations for why Sa... by dragonwriter on Monday November 20, 2023 @11:48PM (#38359391)
> Keep in mind this predates basically ANY kind of source control.

It might be before they were ported to DOS or OS/2, but it definitely wasn't before source control existed (SCCS and RCS were both definitely earlier.)

Re:OpenAI's employees were given two explanations for why Sa... by adastra22 on Tuesday November 21, 2023 @12:18AM (#38359646)
OK: Keep in mind this predates basically ANY kind of source control in common usage in software engineering.
Re:OpenAI's employees were given two explanations for why Sa... by bawolff on Tuesday November 21, 2023 @12:23AM (#38359695)
3x the work may still fall under reasonable cost.

If architectured properly (big if) you can split up the project appropriately so there is a common core and individual parts for specific OS.

Is it extra effort? Sure. Impossible? Definitely not.

Re:OpenAI's employees were given two explanations for why Sa... by WalterBright on Tuesday November 21, 2023 @03:37AM (#38361011)
I've also successfully converted some rather large x86 assembler programs into C, so they could be ported to other platforms. It's quite doable by one person.

(Nobody else wanted the job, but I thought it was fun.)

Re:OpenAI's employees were given two explanations for why Sa... by gruturo on Tuesday November 21, 2023 @06:27AM (#38362229)
Uh? Quite wrong.

SCCS was created in 1973. We're talking about over a decade later.

Also primitive forking, diffing and merging could be (painfully) done even with crude tools, which did exist.

Re:OpenAI's employees were given two explanations for why Sa... by adastra22 on Tuesday November 21, 2023 @02:57PM (#38369170)
These were not in common usage in the software industry.
Re:OpenAI's employees were given two explanations for why Sa... by Dylan16807 on Monday November 20, 2023 @10:15PM (#38358606)
Fork just means two groups start with the same code and work independently.

It was a thing.

Re:OpenAI's employees were given two explanations for why Sa... by jrflowers on Monday November 20, 2023 @10:44PM (#38358856)
Should’ve just made it an Electron app
Re:OpenAI's employees were given two explanations for why Sa... by DonHopkins on Monday November 20, 2023 @10:54PM (#38358945)
You'd need a Beowulf Cluster of Apple //e's to run an Electron app in 1983!
Re:OpenAI's employees were given two explanations for why Sa... by jrflowers on Monday November 20, 2023 @11:00PM (#38358989)
They could have charged a subscription for their cloud based offering then
Re:OpenAI's employees were given two explanations for why Sa... by adastra22 on Monday November 20, 2023 @11:04PM (#38359029)
Been a long time since I heard "Beowulf Cluster"!
Re:OpenAI's employees were given two explanations for why Sa... by inDigiNeous on Tuesday November 21, 2023 @08:49AM (#38363551)
Had to log in just to upvote this comment, brought back nostalgia from the Slashdot days of yore.
Re:OpenAI's employees were given two explanations for why Sa... by saalweachter on Tuesday November 21, 2023 @10:14AM (#38364714)
Eh, C of this era, you're definitely talking some sort of #ifdef PLATFORM solution.
Re:OpenAI's employees were given two explanations for why Sa... by dylan604 on Monday November 20, 2023 @10:27PM (#38358713)
Apple had a skunk works team keeping each new version of their OS to compile on x86 long before the switch. I wonder if the Lotus situation was an influence, or if ensuring your software can be made to work on different hardware is just an obvious play?
Re:OpenAI's employees were given two explanations for why Sa... by mvkel on Tuesday November 21, 2023 @12:41AM (#38359817)
IBM was bankrolling all the development. They only had one choice.
Re:OpenAI's employees were given two explanations for why Sa... by clnq on Monday November 20, 2023 @10:32PM (#38358754)
Consider for a moment: this is what the board of one of the fastest growing companies in the world worries about - kindergarten level drama.

Under them - an organization in partnership with Microsoft, together filled with exceptional software engineers and scientists - experts in their field. All under management by kindergarteners.

I wonder if this is what the staff are thinking right now. It must feel awful if they are.

Re:OpenAI's employees were given two explanations for why Sa... by valine on Monday November 20, 2023 @08:44PM (#38357734)
Steve Jobs famously had two iPhone teams working on concepts in parallel. It was click wheel vs multi-touch. Shockingly the click wheel iPhone lost.
Re:OpenAI's employees were given two explanations for why Sa... by brandall10 on Monday November 20, 2023 @09:59PM (#38358464)
I thought the design team always worked up 3 working prototypes from a set of 10 foam mockups. There was an article from someone with intimate knowledge of Ives lab some years back stating this was protocol for all Apple products.
Re:OpenAI's employees were given two explanations for why Sa... by mikepurvis on Monday November 20, 2023 @10:04PM (#38358508)
Another element of that was the team that tried to adapt iPodOS for iPhone vs Forstall's team that adapted OSX.
Re:OpenAI's employees were given two explanations for why Sa... by kridsdale1 on Monday November 20, 2023 @10:43PM (#38358839)
I think it was also a contest between all web ui (like WebOS) bs Cocoa.
Re:OpenAI's employees were given two explanations for why Sa... by throwawayapples on Monday November 20, 2023 @09:29PM (#38358196)
and the Apple (II etc) vs Mac teams warring with each other.
Re:OpenAI's employees were given two explanations for why Sa... by kmeisthax on Monday November 20, 2023 @10:10PM (#38358567)
You're thinking Lisa vs. Mac. Apple ][ didn't come into the picture until later when some of the engineers started playing around with making a mouse card for the ][.
Re:OpenAI's employees were given two explanations for why Sa... by fakedang on Monday November 20, 2023 @09:47PM (#38358355)
Seriously? Click wheel iPhone lost shockingly? The click wheel on most laptops wears out so fast for me, and the chances of that happening on a smaller phone wheel is just so much higher.
Re:OpenAI's employees were given two explanations for why Sa... by potatoman22 on Monday November 20, 2023 @09:52PM (#38358412)
(It was sarcasm)
Re:OpenAI's employees were given two explanations for why Sa... by fakedang on Tuesday November 21, 2023 @12:20AM (#38359676)
Oops, sorry, didn't get that. I had suspected it was one of those Luddite HNer comments bemoaning changes in tech, and nostalgically reminiscing on older times.
Re:OpenAI's employees were given two explanations for why Sa... by discordance on Monday November 20, 2023 @09:30PM (#38358206)
Happens all the time.

Teams of people at Google work on the same features, only to find out near launch that they lost to another team who had been working on the same thing without their knowledge.

Re:OpenAI's employees were given two explanations for why Sa... by zebnyc on Monday November 20, 2023 @09:56PM (#38358437)
How does that work? Do they have the same the same PM, requirements? Is it just different tech / achitectures adopted by different teams. Fascinating
Re:OpenAI's employees were given two explanations for why Sa... by hanniabu on Monday November 20, 2023 @10:45PM (#38358862)
Give a goal (ex. make it more intuitive/easier for the user to do X), have 2 teams independently work on it, A/B test them, winner gets merged.
Re:OpenAI's employees were given two explanations for why Sa... by discordance on Tuesday November 21, 2023 @09:53AM (#38364385)
It is fascinating, very wasteful and also often devastating for the teams involved who worked very hard who then have their work thrown away.

PMs/TPMs/POs may not know as they're on different teams. Often it's just a VP game and decided on preference or a power play and not on work quality/outcome.

Re:OpenAI's employees were given two explanations for why Sa... by PurelyApplied on Monday November 20, 2023 @09:36PM (#38358268)
Sure, Jan.
Re:OpenAI's employees were given two explanations for why Sa... by fluidcruft on Monday November 20, 2023 @09:38PM (#38358280)
I guess it depends on whether any of them actually got the assignment. One way to interpret it is that nobody is taking that assignment seriously. So depending on what that assignment is and how important that particular assignment is to the board, then it may in fact be a big deal.
Re:OpenAI's employees were given two explanations for why Sa... by kumarvvr on Monday November 20, 2023 @09:41PM (#38358306)
Does a board give an assignment to the CEO or teams?

If the case is that the will of the board is not being fulfilled, then the reasoning is simple. The CEO was told to do something and he has not done it. So, he is ousted. Plain and simple.

This talk about projects given to two teams and what not is nonsense. The board should care if its work is done, not how the work is done. That is the job of the CEO.

Re:OpenAI's employees were given two explanations for why Sa... by fluidcruft on Tuesday November 21, 2023 @10:29AM (#38364927)
Frankly the information that is available is extremely non-specific and open to interpretation and framing by whoever wants to tell one story or another. The way I see it something as specific as "has not done xyz" is a specific thing that can be falsified and invites whatever it is into the public to be argued about and investigated whereas "not sufficiently candid" does not reveal much and just says that a majority of the board doesn't trust him. Altman and all the people directly involved know what's going on, outsiders have no need to know so we're just looking at tea leaves and scraps trying to weave narratives.

And I agree the board should care if the work is actually done and that's where if the CEO seems to be bluffing that the work is being done or blowing it off and humoring them then it becomes a problem about the CEO not respecting the board's direction.

Re:OpenAI's employees were given two explanations for why Sa... by hooloovoo_zoo on Monday November 20, 2023 @10:02PM (#38358493)
Giving two groups of researchers the same problem is guaranteeing one team will scoop the other. Hard to divvy up credit after the fact.
Re:OpenAI's employees were given two explanations for why Sa... by m3kw9 on Monday November 20, 2023 @10:21PM (#38358672)
How did they get 4 board to fire him because he tried to A B test a project?
Re:OpenAI's employees were given two explanations for why Sa... by ldjkfkdsjnv on Monday November 20, 2023 @09:53PM (#38358419)
Also when a project is vital to a company, you cannot just give it to one team. You need to derisk
Re:OpenAI's employees were given two explanations for why Sa... by quickthrower2 on Monday November 20, 2023 @09:40PM (#38358296)
Was that verbatim the reason or an angry persons characterisation?