The New Inflection

https://inflection.ai/the-new-inflection

by break_the_bank • 2 years ago

194 157 2 years ago

inflection.ai

fnbr• 2 years ago

Wow. This is surprising. Not even an acquihire. I’m very curious about what happened internally to lead to this.

riku_iki2• 2 years ago

no more money left..

tfehring• 2 years ago

It would be extremely poor form for founders to intentionally leave their startup at this stage to take another job. In all likelihood they were pushed out by the board.

My guess is that the founders were all in on Pi, their conversational language model, and investors lost confidence in their ability to build a sustainable venture-scale company on the ~4th best AI chatbot.

GaggiX5• 2 years ago

I've been using their new model for a few days and it works surprisingly well, it's also free and works in the EU.

I don't know how I feel about Microsoft's involvement here.

m_ke2• 2 years ago

No traction, was probably clear they didn't have a shot at growing the consumer side.

GaggiX5• 2 years ago

They have +1 mln daily users.

rvz2• 2 years ago

It appears that Inflection AI made no sense to begin with and Pi was quite frankly a performative research demo and didn't generate enough money for the VCs to justify another fundraising round. How is Inflection AI worth $4BN?

What can Pi do that is unique over the best of cloud LLMs and the hundreds of $0 free LLMs out there?

It appears that it is a vehicle for VCs to quickly run this company to the ground for a quick exit, knowing that this company is extremely overvalued.

Probably after this acqui-hire, the value of Inflection AI is now down to its real value of $200M at most.

telltruth3• 2 years ago

My guess is that Mustafa wanted to sell to MSFT at 10X but MSFT didn’t wanted pay that kind of money. Mustafa was ok with fire sale but VCs were greedy. Mustafa then quite in rage.

kcorbitt• 2 years ago

So sounds like the real news is that Microsoft basically acquihired the Inflection founding team?

nexuist• 2 years ago

I hope one day we can drop the corporate facade so these blog posts can be two to three sentences of "Our company sucked. Microsoft wanted to give me a fuck ton of money. I decided to take it. Some other guy is in charge now. Good luck!"

gcnnbdff• 2 years ago

A lot of people don’t know this but all these VC valuations are fake, they had this planned years ago to fake growth.

f6v• 2 years ago

Maybe someone try training an LLM to do just that.

bee_rider2• 2 years ago

Does getting acquhired by MS imply your company sucked?

There are things I like about where I’m at but if MS really wanted to they could send enough dump trucks full of cash to ruin the local traffic situation. Is it worth staying at a place if you can’t even get out of the driveway to get to your favorite little cafe?

aeternum• 2 years ago

Ironically, AI will probably help to provide this.

Could probably do it now with a browser plugin and the right llm prompt. Perhaps call it the un-BSifier.

kookamamie• 2 years ago

Hear, hear! There is a market for a no-bs-happenings.com.

duxup• 2 years ago

I think we can assume almost all acquisitions are "this cash is better than continuing to operate as we were.... make of that what you will".

guytv• 2 years ago

I asked ChatGPT to summarize, this is what it came up with:

* Inflection shifts focus to AI studio business, developing custom AI models for commercial use.

* Plans to host Inflection-2.5 on Microsoft Azure, with plans for other cloud platforms.

* API launch soon, sign-ups for early access open.

* Co-founders Mustafa and Karén to start Microsoft AI, leaving Inflection.

* Sean White appointed new CEO; Reid Hoffman remains on board.

* No immediate changes to Pi service; privacy and data policies unchanged.

rllearneratwork2• 2 years ago

is it really a win on MSFT part though?

vineyardmike2• 2 years ago

Absolutely.

Pi (their chat bot) is pretty nice to talk to. It’s really good with the whole para-social aspect of chatting without being weird. Is that useful? Maybe for research, probably less so for a direct product.Microsoft can probably get something useful from that.

More importantly though, Microsoft is now hosting a whole suite of LLMs on Azure. This is the lesson they learned after OpenAI had that acute leadership crisis. This is another hedge against OpenAI. I wouldn’t be surprised to see Microsoft start distancing themselves from OpenAI in the future.

zooq_ai• 2 years ago

It is not.

Mustafa Suleyman is the ultimate AI grifter.

He has ZERO technical skills. ZERO hardcore STEM background. He was just friends with Demis at the right time and road that hype train.

At Google he had a similar role VP of AI products or something and he contributed nothing (except all the garbage ethics, safety crap that didn't help google, probably kneecapped it actually)

This is Satya Nadella's worst AI move yet.

telltruth3• 2 years ago

They aquihired 3 people who didn’t do any technical work. Nadella miscalculated big time here.

jonny_eh• 2 years ago

Not if it kneecaps a key competitor.

ipsum2• 2 years ago

Fortune reported that Microsoft hired most employees from Inflection.

m_ke2• 2 years ago

Right, Karen Simonyan is not technical at all...

telltruth3• 2 years ago

They had raised massive amount and not from good patient investors. No traction means Mustafa got fired. This is not surprising though but what is surprising is MSFT picked him up. The guy is not technical, is not even visionary and had just got lucky hanging out with Demis. I would think Satya had better taste.

toomuchtodo• 2 years ago

Never underestimate the power of soft/social skills.

rqtwteye3• 2 years ago

That’s where the real money is. Even the best technologists won’t make it very far if they don’t know how to play politics.

skepticATX• 2 years ago

Mustafa wasn’t exactly on his best behavior at DeepMind, if the allegations are to be believed. It really is surprising that Microsoft would hire him.

refulgentis3• 2 years ago

It's crazy there's zero accountability for bad behavior in tech. I went through my own story at Google, and seeing them say it was vaguely bad before promoting him to VP mirrors exactly the "intervention" I saw.

The deck is completely stacked against you based on hierarchy. Behavior that a fast food manager would proactively solve in 30 seconds gets ignored in white collar tech. No one above you will even mention it - they know you can't win and they just hope you'll quietly give up.

If someone above you in the informal hierarchy is messing with you, there's massive confirmation bias if you complain. They'll spin it to whoever you complain to make you the bad guy. HR never helps - their job is to investigate, and then give the results to someone 2-3 steps above you to do something with.

The higher ups control the outcome, and they designed the power structure in the first place, their confirmation bias is accept the spin.

If you want to survive, avoid conflict 100% of the time. Let people blame you, fail reviews undeservedly.

My Google career ended from just doing exactly what I was supposed to do in order to get a 3 year delayed project done, that 4 separate VPs had been asking for all those years. I spent 6 months warning my manager fuckery was afoot. Didn't matter. TPM witnessed and defended me, didn't matter. Guy who led it hired his unqualified childhood buddy to replace me. Didn't matter. All on me. Everyone wanted to do it, and gee whillakers, refulgentis went mad and dropped the ball completely for some reason.

Of course, 6 months later they delayed the project a 4th year because they could, documenting the only downside being a strained relationship with a less influential partner team. (my orgs managers didn't realize their...unvarnished...takes were in a doc shared with all of Google)

At the end of the day, HR will funnel you into taking mental health leave -- 6 months worth, exactly long enough that an EEOC complaint can no longer be filed. (took me 6 years to realize why "disgruntled Google employee" news articles always included a bit referencing leave/6 months off as if it was a bad thing. go/mh-leave if you're at Google. You don't actually need to talk to HR, and I don't recommend going to them ever. I didn't for this, but they wouldn't have helped.)

The whole system is broken.

Bjorkbat• 2 years ago

He also left DeepMind because of allegations of bullying employees (https://en.wikipedia.org/wiki/Mustafa_Suleyman). Between that, what you just brought up, and the strange PR blitz he went on to promote his book, I kind of predicted Inflection would run into major trouble well before OpenAI and Anthropic

Workaccount2• 2 years ago

Some of the most damaging people I have met are people that you are absolutely enthralled with when you first get to know them.

spaceman_2020• 2 years ago

True, and some of the most effective people I know haven’t been impressive when you first meet them.

I’m now extremely wary of blusteringly confident, glib talkers. Experienced a high correlation between these traits and sociopathy.

rqtwteye3• 2 years ago

That would apply probably to most famous entrepreneurs. Whatever I have read about Jobs, Musk, Gates and others, they all are very willing to abuse people to achieve their goals.

coffeebeqn2• 2 years ago

First impressions is a sales skill

baq2• 2 years ago

There’s a whole extremely famous book series on this very topic, you might have heard about it. The first volume’s title is ‘Paul is bad’ but it’s more widely known as ‘Dune’.

vouaobrasil6• 2 years ago

As part of this, we’re thrilled to announce that we will now host Inflection-2.5 on Microsoft Azure helping us get it into the hands of creators everywhere.

Hahah. You know, the word 'creator' has truly become co-opted by consumerism and greedy bigcorps. What a lot of posts like this never mention is one of the primary uses of AI: to make more efficient the system of trying to get us to buy even more things we don't need. I mean, who are we kidding? In a well-balanced life, we shouldn't need or even interact with personal AIs. We should slow down and appreciate what we have and seek for simplicity.

If there's too much data to be handled in business, it means business is not going in the right direction, not that we need new tools to handle it.

If we are pressured to write more, then we are creating things not of true value but merely to amuse.

If we feel like becoming more efficient is a good thing, it simply is an extension of the original psychological manipulation of advertising demanding MORE for the industrial machine.

Let's not fool ourselves into thinking this garbage is something good.

airstrike4• 2 years ago

It's hard to agree with most of what you said without starting from the big assumption that "capitalism is bad" and "being anti-capitalism is a virtue"

Becoming more effective in and of itself simply means becoming more productive, which means creating goods and services with less, which means more "welfare" to people

If people are being convinced to buy shit they don't need, that's a separate issue. But don't throw the baby out with the bathwater

In my well balanced life, I love interacting with LLMs. I also appreciate the things I have all of the time, without needing to slowing down technological progress to do so

vouaobrasil6• 2 years ago

In my well balanced life, I love interacting with LLMs. I also appreciate the things I have all of the time, without needing to slowing down technological progress to do so

Okay that's fine, but you are not the rest of the world. While you enjoy the ultimate fruits of capitalism, there are people in developing countries suffering because of it through a lack of freedom and a destructiom of the landbases that they could once depend on. When climate change comes because of consumerism, it is you and other well-off people that will be able to move first.

OF COURSE, you will disagree with me, because you are probably at the apex. (Just the very fact that you have a well-balanced life means you are close to it, proportionally speaking to the entire human population.)

I don't claim that all capitalism is bad BTW. Rather, what I claim to be bad is global capitalism where other considerations have been eliminated or decimated.

ggpsv2• 2 years ago

I find it curious that your reaction to this is to reduce OP's point to the "capitalism is bad" trope when it is in fact more nuanced than that. His point isn't anti-capitalist per se - it critiques the traits of the dominant economic system that many of us currently live in.

I won't assume bad faith on your part, but your argument also relies on a big assumption with regards to productivity gains, and who reaps them, and at the detriment of whom and what.

BriggyDwiggs422• 2 years ago

Pretty hard agree here, but I think there’s space for a gripe.

Our default state as humans is pretty awful and materially deprived, so I think making technology to improve our conditions is good. Making technology that does powerful things in the real world is difficult partly because of the vast quantity of data in its structure. If we make powerful AI systems equipped to handle that scale and built to make our lives better, it would be good, but we’re failing to do that which is the issue.

vouaobrasil6• 2 years ago

Our default state as humans is pretty awful and materially deprived, so I think making technology to improve our conditions is good.

It's not a dichotomy! When I rail against technology, I am not saying that everyone need go back to the stone age. But don't you think there might be a happy medium?

At least personally, I've found that giving up a fair amount of technology in many cases has actually improved life in ways that I didn't think possible. Not everyone needs to be ascetic, but we seem to be headed to a life very tightly integrated with technology, and I think it's arguable that THIS POINT is past the point of that happy medium.

ChildOfChaos• 2 years ago

Wouldn't a personal AI help you live a more well-balanced life, allowing us to offload tasks to it so we can appreciate the world around us and simplicity, rather having to do it all ourselves?

vouaobrasil6• 2 years ago

So far it seems that other automations have not done so. We've offloaded many tasks to computers, but now we spend more time than ever in front of a screen (see https://www.statista.com/statistics/645644/north-america-dai... for example). More people live in cities, with less opportunities for fresh air. And let's not forget that the system itself is so complex now with so many dependencies that we surround ourselves with anything but simplicity.

There might be a world where what you say is possible of course, but I don't see it in this one when it comes to more advanced forms of automation.

In fact, sometimes the joys of life can be found in the most simple of mechanical tasks. Of course, not everyone thinks so.

crooked-v3• 2 years ago

The problem is that none of them are actually good enough for that yet. The closest I've seen are calendar-managing LLM services, and even those only really work because it's a tightly-constrained task set and easy to confirm the results as good or not.

Edit: I still can't even trust Siri to be fully accurate 100% of the time about turning the lights on and off, let alone anything more complicated than that.

jjtheblunt• 2 years ago

Why not just simplify, and not need such?

Aerbil3132• 2 years ago

The only inflection I see here is Microsoft single-handedly dominating global tech industry and its future after its years in sleep. US DoD must have been really worried about the US Empire’s continued dominance.

unethical_ban2• 2 years ago

Uh-huh.

Amazon, Google, Meta, and Apple are has-beens.

A plausible theory with less coordination required: Microsoft is a large company with lots of money attempting to regain competitive advantages against their rivals through acquisition.

CuriouslyC2• 2 years ago

While that's a true story today, Satya is brutally outplaying Google, Amazon is a utility company and Apple's domain is hardware and will likely continue to be. Mark is making some good strategic plays to keep the field even, but Microsoft really is poised to clean up as AI takes off.

Aerbil3132• 2 years ago

Surely the others have ties with Pentagon too. But Microsoft is the biggest company in the world and the one most intertwined with the US government.

bugglebeetle• 2 years ago

Between this and the Mistral deal (that’s currently under investigation by the EU), Microsoft looks to be really trying to get back into the monopoly business with AI.

baq2• 2 years ago

he who controls the spaice, controls the universe, after all.

thomastjeffery• 2 years ago

Microsoft has been in the monopoly business since at least 2001.

darth_avocado2• 2 years ago

Is tech hiring great? No. Is it the absolute worse it can be? Also no. Internship or apprenticeship instead of tech interviews are just take home tests that last multiple months, and that is a huge waste of time.

Yes, broken interview system currently rewards people who interview more, but when you combine it with other signals like resume, project presentations and other things, the chances of hiring bad candidates are low. Yes, you miss out on a lot of great candidates, but that is not a problem that big tech companies need to solve. Besides, not knowing the status of your employment 3-6 months down the line is not a great thing for candidates.

Analog24• 2 years ago

Internships are definitely not a waste of time. First of all, they pay well in and of themselves (at least in tech). Second of all, most internships are filled by students on their summer break. What better use of that time than getting an inside view of a company they might want to work for? From the companies perspective it gives them a much better idea of how well the candidate performs and how they fit in at the company, giving much higher confidence in a hiring decision that could lead to significant future impact.

darth_avocado2• 2 years ago

I think you misunderstood. Internships for students during the summer definitely makes sense. However, i was talking about internships/probation as a way to evaluate candidates instead of interviews. Meaning if you are a staff engineer with 10+ years of experience, you’re still going to be hired conditionally for 3-6 months to see if you’re good enough for the team. I personally find that very prone to abuse.

bcopa• 2 years ago

Pretty mind-boggling how ChatGPT wrapper apps built by dev houses in Poland are absolutely crushing it, while Pi failed to gain any traction despite raising billions.

BoorishBears• 2 years ago

More mind-boggling they raised billions to do something any LLM can do with some prompts (be more personable)

Pandabob• 2 years ago

Have not heard of these. Where can I learn more about them?

toisanji• 2 years ago

examples please

dbish3• 2 years ago

Which?

lumost• 2 years ago

The interesting emphasis to me is commercial customers. They are acknowledging that the competition to be the nth gpt vendor is too stiff for them, and they aren’t successful as an independent venture funded research lab.

I can think of 4-10 other large vc funded operations in this boat.

lvl102• 2 years ago

This space is moving so fast and it’s hard to keep track of SotA. How are management even keeping track? I suppose MSFT has enough cash to buy ‘em all.

coffeebeqn2• 2 years ago

State of the art is still GPT-4? Others are playing catch up or hitting very similar benchmarks

andy99• 2 years ago

Is Microsoft AI the thing that Altman was going to be in charge of in the short lived agreement he had with MS?

mupuff1234• 2 years ago

Wasn't Mustafa kicked out of deepmind/Google due to harassment allegations? Seems like a pretty pathetic move from MSFT.

Animats• 2 years ago

Ah, PR. Buzzword buzzword buzzword ... second screen ... "two of our three co-founders ... will be leaving Inflection to start Microsoft AI",

gwern2• 2 years ago

Who gets the GPU clusters...?

ShamelessC2• 2 years ago

Do they own their GPU clusters?

gwern2• 2 years ago

You don't train GPT-3+ models or serve millions of users off Amazon spot instances, that's for sure! And whatever arrangement they have is likely transferrable or salable.

nabakin2• 2 years ago

I have some serious issues with this company's PR.

They say their Inflection-2.5 model is the world's best personal AI[0] which is a dumb claim to make considering it is done off of automated benchmarks which we know are flawed and even if you assume automated benchmarks are good enough to claim that title, it would be held by dozens of other open weight models on HF, not Inflection-2.5.

They said their Inflection-2 model was the second best in the world[1] while comparing it to Palm-2 which no one considers close to the best in the world. They again, based their claim on automated benchmarks which anyone knowledgeable in the space would know can be gamed and is not representative of actual conversational performance. (take a look at the Lmsys Arena Leaderboard for a better metric)

They list other models they consider good while failing to mention or compare to Mixtral 8x7b, the best open model that exists.

And they introduce buzzwords no one in the area uses like IQ and EQ as if they are innovative concepts.

Making big, bold claims without evidence is the exact kind of manipulative PR speak I'd expect from a company with little to no substance.

[0] https://inflection.ai/inflection-2-5

[1] https://inflection.ai/inflection-2

posix86• 2 years ago

Agreed, read a few sentence & move on.

maest2• 2 years ago

Fwiw I wasn't familiar with them and found gp's comment useful. I actually upvoted

nabakin2• 2 years ago

I think it's good to raise awareness of bad practices if you recognize them.

Leary• 2 years ago

Exactly, honestly I think their business model could work if they have a GPT-4 level model for subscription.

RaftPeople• 2 years ago

"After just 3 days of being incorporated and just 2 days of actual work performed, we have now created the best AI on the planet with trillions of active subscribers, per second.

The quality of our product is so high, it is being compared to a perfect diamond the size of the entire solar system. But, we are humble, and while we think our tech is pretty darn good, we know that we can do better. That's why we are introducing v2 in just 2 business days from today, which will render all other forms of intelligence (human and artificial) irrelevant."

disqard• 2 years ago

I wish I could boost your comment even higher.

It really captures what I felt as I started to read this PR bullshit. I fed it through an LLM to summarize it, and there was no substantive content in that summary.

Lampooning aside, the sobering reality is that a tiny number of people are acquiring O($10^8) wealth (or more) via such shenanigans, and that is the reason we will see a lot more of this.

fullstackchris• 2 years ago

Tell me about it. I'm really starting to get ticked off trying to wade through garbage hype and actual improvements across everything "AI"

bevekspldnw• 2 years ago

I just skimmed and came away with “Satya is dumping more Azure credits on some AI thing to screw with Big G again”. The rest was noise.

petre3• 2 years ago

Dunno, it's pretty damn good and quite useful. I asked Mistral 7b and Claude Opus to count the banned substances on the WADA anti doping list. Both avoided a straight response giving back loads of bs. Claude was next to useless. Mistral gave a somewhat acceptable answer counting substances on every chapter after much persuasion. Inflection's pi AI gave me a straight answer to the first question without any bs. I like its casual tone and the fact that it does not use needlessly superfluous language like the GPT models.

crooked-v3• 2 years ago

Also, and I'm going to put this real bluntly, a "personal AI" isn't worth crap to a lot of people until it's allowed to talk about sex.

I'm not even talking for porn purposes. Sex is a basic and healthy part of most peoples' lives, any number of relationship issues revolve around it, and for plenty of people it's literally their profession, legal or not. Anything that pretends it can 'personally' help a broad spectrum of people while treating sex as verboten is just bullshit.

echelon• 2 years ago

A lot of people, on every side of every aisle, think that censorship of opinions and ideas that they do not hold is A-OK.

And here we are ensuring our models are cut in the same cloth we are. Reflections of the cages we wish to design for others.

War is peace. Oppression is safety.

smoyer• 2 years ago

Their web site is horrible on Mobile too!

atoav• 2 years ago

Well the NFT-hype people have to go somewhere..

They learned that most people can't tell the difference between actual technical concepts and PR speak and it worked to sell NFTs, so now that AI is all the hype, guess what.

ChrisArchitect• 2 years ago

The New Inflection

Related Stories