Francis Yates

649 posts

Francis Yates banner
Francis Yates

Francis Yates

@considerahead

Senior Deep Learning Research Scientist at @3M Dublin. Engineer with a passion for robotics. I work on 3D mesh Transformers. All views are mine

Dublin City, Ireland or UK Katılım Ocak 2016
931 Takip Edilen161 Takipçiler
Francis Yates retweetledi
Yann LeCun
Yann LeCun@ylecun·
Most of these signatories have a distorted view of what is coming next with AI. The distortion is due to their inexperience, naïveté on how difficult the next steps in AI will be, wild overestimates of their employer's lead and their ability to make fast progress, and financial incentives to hinder open source AI platforms (since almost all of them work for providers of proprietary AI systems). As you and I know, generations after generations of AI researchers have made the same mistake: thinking that human-level AI is "just around the corner" because of some new paradigm they are working on. They have consistently underestimated how difficult it is because they have an incentive to dismiss the limitations of their favorite paradigm as mere "engineering problems," and they can't foresee the obstacles they'll bump into before bumping onto them. We are making progress towards human-level AI, but we're still far from it. Some of us may have vague roadmaps for how to get there. But no one has a credible blueprint, let alone a demonstrable prototype. Before such a prototype exists (perhaps with the learning abilities of a house cat), regulating AI R&D because of a fear of existential risk is highly premature. There is nothing wrong with sensible legislation to regulate the *deployment* of AI application. But legislation that hinders open research and open source AI platforms, and that makes open source AI developers liable for what people do with their code is extremely regressive.
English
75
131
1.2K
101.1K
Feljin Jose
Feljin Jose@Feljin_J·
Fair play to this guy for recognising the great work done by the Greens. That's a decent bus lane and a fully segregated cycle lane right there. Now if Eamon could do something about the cars in the bus lane..
Feljin Jose tweet media
English
61
30
600
73.6K
Francis Yates
Francis Yates@considerahead·
@LukeDaly1979 @Feljin_J The section from 5 lamps to town also makes the commute safer for cyclists, scooters, etc. So it’s not exclusive to Clontarf commuters.
English
1
0
0
31
Lukus
Lukus@LukeDaly1979·
@considerahead @Feljin_J There has always been traffic there in the morning.But now there is traffic there all day every day. Also the people of Summerhill have been subjected to a tripling of traffic along Ballybough road, but as usual,working class communities don't matter as much as Clontarf cyclists
English
1
0
0
104
Francis Yates
Francis Yates@considerahead·
@LukeDaly1979 @Feljin_J When I was a student, I cycled that route everyday, saving 3 euro ish per day on the bus. It was like 5km, so too far to walk in the morning. Cycling was faster than the bus too. The extra 20 quid per week made a difference. The car traffic was exactly the same back then, too.
English
1
0
0
18
Lukus
Lukus@LukeDaly1979·
@Feljin_J Yeah, creating significant hardships for hard pressed working families, because they have no option but to drive, is really great work. The Greens are a classist party. Why do you hate poor people so much?
English
1
0
4
176
Francis Yates
Francis Yates@considerahead·
@gallabytes @ESYudkowsky @ylecun In response to this thread, the critical first try is something unbelievably rare in engineering. Almost all progress in the 20th/21st has been iterative and deliberate. Seldom do we we have light switch moments as @ESYudkowsky suggests. Even in AI field, we've had almost none
English
0
0
3
92
theseriousadult
theseriousadult@gallabytes·
@ESYudkowsky @ylecun LeCun clearly disagrees with you about how critical the first try is. He expects a series of engineering problems which do not suddenly get more difficult. You expect a sudden phase change requiring us to solve philosophy ahead of time to make it through.
English
4
0
15
1.1K
Yann LeCun
Yann LeCun@ylecun·
Technologies that empower humans by increasing communication, knowledge, and effective intelligence always cause opposition from people, governments, or institutions who want control and fear other humans. The arguments against open source AI today mirror older arguments against social media, the internet, printers, the PC, photocopiers, public education, the printing press,.... Historically, those who have banned or limited access to these things have not been believers in democracy.
Nora Belrose@norabelrose

The terrorism argument against open source AI also applies to anything that increases the effective intelligence of humans: the internet, public education, nutrition, etc. It's a fully general argument against human empowerment.

English
97
221
1.3K
574.6K
Francis Yates retweetledi
Brian Mac Mághnais
Brian Mac Mághnais@TheBrianMcManus·
I try my best to not talk about Irish politics, but getting sent things like this by my Garda friends just angers me too much. This happened today. The police (Gardai) in Ireland have been attacked and bullied by scumbags for years now. They know they can get away with it. Stripped of any power to do anything about it. The Gardai are trained to retreat and not engage when violence breaks out. This is what happens when you remove your police forces ability to police. The Gardai voted with an enormous majority vote of no confidence against the head of the Gardai. Our minister for justice ignored them. Overworked, underpayed, and stripped of any protection in policing. I know policing in America is a sore topic for good reason. Ye are on the other side of the spectrum. Too much power. This is what happens when you swing to the other side. Chaos. I’m ashamed of how Ireland is being governed. A failed socialist state. It’s why I left, and it’s why I won’t be going back any time soon. Helen McEntee, you need to resign.
English
21
26
341
120.6K
Francis Yates
Francis Yates@considerahead·
@billpeeb Great talk at ICCV! Such a interesting architecture. 🤗
English
0
0
0
48
Francis Yates
Francis Yates@considerahead·
@dara_moran @crazyhouseprice So people that rent have to rent from private landlords? Private landlords in my experience are worse than renting from a company.
English
0
0
0
17
Dara
Dara@dara_moran·
@crazyhouseprice Or a ban on corporate ownership of residential property
English
1
0
2
539
Ciarán - Crazy House Prices
Ciarán - Crazy House Prices@crazyhouseprice·
This would be the worry, but easily fixed by implementing a ban on all bulk purchases of already-built homes by investment funds. If I was the leader of a government being hammered in the polls, it's certainly something I would look at!
Frank McDonald (@frankmcdonald60.bsky.social)@frankmcdonald60

Good to see this, but I'd be wary of developers' pledges at the planning stage. Cairn Homes developed Griffith Wood in Marino as a build-to-sell scheme, but sold all 342 apartments to US vulture fund Greystar in March 2022 for €176.5m. Now it's exclusively Build-To-Rent ...

English
2
3
42
14.7K
Cormac McKay
Cormac McKay@cormacDublin·
@DublinCommuters The start stop nature of double decker bus travel and now a hub and spoke model under BusCONnects will not speed up journey times never have never will! rte.ie/archives/2023/…
Cormac McKay tweet media
Dublin City, Ireland 🇮🇪 English
1
0
1
154
Dublin Commuters
Dublin Commuters@DublinCommuters·
“We have a finite amount of space in cities. The only way to make sure buses get through Dublin city centre quickly is to allocate spaces to buses. That will mean restrictions on cars.. that space has to come from somewhere.” newstalk.com/news/newstalk-…
English
5
0
48
3.7K
Francis Yates
Francis Yates@considerahead·
@AlphaSignalAI Or the progress is due to an increasing number of people moving into the field?
English
0
0
0
9
Francis Yates
Francis Yates@considerahead·
@marktenenholtz For a thought experiment, take the weights run a few epochs fine tuning on a slight different dataset. are the weights still non-commercial? You might get 99% of the performance but are the weights considered the same?
English
0
0
0
34
Mark Tenenholtz
Mark Tenenholtz@marktenenholtz·
Sure, the code is released under GPLv3. But the weights ARE NOT. They’re strictly non-commercial.
Djamé..@zehavoc

@marktenenholtz it's was a rhetorical question. of course it allows commercial usage if you're willing to share derivatives work under the same conditions. GPL3 is to prevent parasitism.

English
6
1
64
26.8K
Mark Tenenholtz
Mark Tenenholtz@marktenenholtz·
Meta released LLaMa yesterday, which outperforms GPT-3 on many tasks with <10% of the parameters! Except…
Mark Tenenholtz tweet media
English
11
107
792
246.2K
Francis Yates
Francis Yates@considerahead·
@wightmanr When you’re training multiple models or doing hyper parameter search, what libraries do you use to track and manage these training runs?
English
0
0
1
165
Ross Wightman
Ross Wightman@wightmanr·
Over the past month I've been working on training ConvNeXt CLIP models with OpenCLIP. To date training efforts have focused on the ViT arch. The original OpenAI paper included modified ResNet models but the results weren't inspiring. ConvNeXt is working amazingly (no mods)!
English
5
28
280
80.5K
Francis Yates
Francis Yates@considerahead·
@tunguz I use both professionally and I’m still waiting to see what PyTorch does significantly ... although I do love the hugging face integration of PyTorch.
English
0
0
0
36
Bojan Tunguz
Bojan Tunguz@tunguz·
So ... do people still use TensorFlow?
English
178
48
1.3K
381.6K
Francis Yates
Francis Yates@considerahead·
@NannaInie Depends if you talking about the model GPT3 or the trained on language GPT3. plenty of uses for the architecture.
English
0
0
1
0
Nanna Inie
Nanna Inie@NannaInie·
Deadly serious, is there ANYONE who can tell me a single good use case or business model for ChatGPT - or even GPT-3?
English
292
46
469
0
Francis Yates retweetledi
Louis Kirsch
Louis Kirsch@LouisKirschAI·
Emergent in-context learning with Transformers is exciting! But what is necessary to make neural nets implement general-purpose in-context learning? 2^14 tasks, a large model + memory, and initial memorization to aid generalization. Full paper arxiv.org/abs/2212.04458 🧵👇(1/9)
Louis Kirsch tweet media
English
7
83
389
0
Francis Yates
Francis Yates@considerahead·
Next week I'm going to be joining @3M as a Senior Research Scientist for AI in dental applications. Can't wait, sad to be leaving @VisualAIPeople after 4 great years.
English
0
0
4
0
Francis Yates
Francis Yates@considerahead·
@crazyhouseprice Taxed such that sale is forced. The price would be right at some point. If they can define RPZ they could do for derelict building too.
English
0
0
1
0
Ciarán - Crazy House Prices
Ciarán - Crazy House Prices@crazyhouseprice·
Somebody owns this beautiful building and allowed it fall into this state of ruin. That, in my opinion, should be a crime. It's up for €250k, and as far as I know, has been sale-agreed a few times over the years. daft.ie/for-sale/terra…
Ciarán - Crazy House Prices tweet mediaCiarán - Crazy House Prices tweet mediaCiarán - Crazy House Prices tweet mediaCiarán - Crazy House Prices tweet media
English
6
6
101
0
Francis Yates
Francis Yates@considerahead·
@sh_reya A couple years ago Tensorflow changed a kwarg in a dropout layer from 'drop_probablity' to 'keep_probability' Suddenly overnight all my models trained without error but didn't perform. I was baffled for days. Positional args 0/10 do not recommend.
English
0
0
1
0
Shreya Shankar
Shreya Shankar@sh_reya·
1+ year after initial release i just now found a bug in my feature generation code where i added epsilon to a denominator and epsilon was 1e7 instead of 1e-7. yay silent errors 🙃
English
10
3
273
0