deepnet

Arguably a win for the community of users in the long term as those developers who would like a safe environment, more akin to a professional one, will be encouraged to contribute.

In the short term Ken will have to wait for the time out to end, so a loss for sure.

On balance the future is bigger than the present so suffering today for a better future is utilitarian.

And there are a lot of sensitive folk who have been scared off of communities by what they saw. In this thread a couple have commented about the pre CoC community, that they would have contributed but for the fear of getting barked at.

As a younger dev I bailed on some IRC communities after getting weirdly threatened.

Frustration after being very patient with one’s donated time and feeling unreciprocated can lead to frayed nerves and in the short term being rude is brief and often effective - so it is very understandable that this sort of thing will arise.

Good to discuss how we handle these difficult situations of onboarding newbies without scaring some of the lurking next generation of devs.

As a headline this penalty seems harsh, but efforts were made to mediate before the committee made the ruling and proved unsuccessful.

This process is defining the future culture of the kernel code community and it is widely accepted that it should be less toxic and more welcoming and akin to what is expected in a professional context where most devs also function and also have to patiently onboard new developers.

There may be room for improvement of the process, for instance, like in many settings the burden of patient onboarding may need to fall on a community process with an onboarding ramp that doesn’t fall only on the shoulders of those accepting patches.

As a community it behooves us to step in earlier and help new devs polish their contributions before submission.

Especially if we see a fraustration brewing, step in and offer code review to the new developers, who want to contribute, but need some guidance.

Often just guidance about the right questions to ask and where and how to find the answers oneself before asking.

The eternal September doesn’t have to be so, offering to help with onboarding a new dev helps everyone.

Spin glasses are a proposed model of learning in Neural Nets[0] along with state changes and sand pile collapses.

So this paper is worth reading if you study learning in neural nets

[0] Spin-glass models of neural networks Amit et al 1985 https://journals.aps.org/pra/abstract/10.1103/PhysRevA.32.10...

This was why Hopfield was awarded a physics Nobel - spin glasses as emergent learning.

Along with Expected Energy of Brownian motion in gases for Hinton’s Boltzmann Machines ( and of course backprop being Newtonian downhill on the manifold of correctness ).

Langland’s worthy crossovers and the utility of the generalisation.

Jinlian (1964–1982) of China was 8 feet, 1 inch (2.46 centimeters) when she died, making her the tallest woman ever. According to Guinness World Records, Zeng is the only woman to have passed 8 feet (about 2.44 meters)

Mean from article 163.

So the facts check out.

Author is correct.

Also very interesting the suggestion that human height is not Gaussian.

Snip :

“ Why female soldiers only? If we were to mix male and female soldiers, we would get a distribution with two peaks, which would not be Gaussian.

“

Which begs the question what other human statistics are non Gaussian if sexes are mixed and does this apply to other strong differentiators like historical time, nutrition, neural tribes ?

Statistics is highly non-trivial. “

Matching risk profiles with business partners or co-founders is vital and often overlooked

Mostly Folk match expectations, I.e. upsides

Downside fear is a very great motivator

A lack of consequential downsides can make folk sensibly choose to cut their losses and move on.

Sometimes this is invisible as parents bailing you out is not known before and may be a private discussion.

All or nothing downsides can mean folk will work themselves close to death to win ( or just not lose )

It also effects how much someone will gamble to get a unicorn rather than settle for a comfy income

Have the downside discussion early on

Discriminatory Snip : “ that she has autism spectrum disorder and writes in a formulaic manner that might be mistakenly seen as AI-generated“

This strikes a chord !

As a former Alumni of Hinton’s MOOC I can attest to his humble, kind nature and clarity, & erudition as a teacher.

Modelling learning as entropy, and heat in his Boltzmann machines was genius as was the 1980s backdrop paper.

Geoff tirelessly evangelised neural nets and machine learning right from his 1970s phd days at Edinburgh.

Despite being in the academic wilderness during the many decades of symbolic AI.

Moore’s law ( & parallel processing via GPUs ) finally caught up with Geoff’s vision and proved him right.

Well deserved !

[post] tl;dr

A ( linked ) list is a collection of CONS ( Construct ) cells.

A CONS cells can extract, via the commands:

‘car’ ( Contents of Address Register ) - the payload ( data item in list );

‘cdr’ ( Contents of Decrement Register ) - the link ( to the next list item, e.g. CONS cell )

The names are for historical reasons ( using two halves of a single register ) lost in the beards of yore.

The linked list appears simple, this belies the fact of its exquisite perfection and the great thought that went into inventing, perfecting and implementing it.

“[The ‘list’ command]… creates a series of conses and sets the CAR pointers to point to each argument. which you have provided. (LIST (LIST 1 2) (LIST 3 4)) => ((1 2) (3 4))

This is brilliant. I currently do this in org mode, pasting bookmarks in long indented trees of research.

This is that but next level, many thanks for sharing your work.

Following rabbit holes, if the trail is preserved, turns my ADHD from inconvenient distraction into a research superpower.

I am intrigued. Please can you link to more about your project and its inspiration. Or write on it, do you have a blog or social to follow ?

Sounds great, please so more - and tell others.

This is lovely work Kleene would be * happy. Rosie is elegant, and thus easy to (human) parse.

Rosie is smooth; smooth is flow - I will use this lots :-D

6.23 GMT spacewalk commences

https://www.spacex.com/launches/mission/?missionId=polarisda...

Must be more than 10 people watching sxhitter livestream capability is hopelessly broke in the UK - SpaceX was streaming well but now they have a redirect to x

Please post alt streams or I guess we’ll just have to wait for youtube highlights

Guess Musky fired too many server guys

Spacewalk cancelled…

Because there is no airlock in the SpaceX Crew Dragon spacecraft used for the flight, the only way to perform a spacewalk is to let all the air out of the spacecraft, and then open one of the hatches. That requires all four crew members to wear spacesuits. NASA and Soviet astronauts conducted spacewalks in a similar manner in the 1960s.

As the Crew Dragon swings around Earth in an elliptical orbit that swings as close as 120 miles to the surface and as high as 460 miles, Mr. Isaacman and Sarah Gillis, a SpaceX engineer, will exit the capsule for about 15 to 20 minutes each.

They will pass through the hatch at the top of the Crew Dragon with the help of a handrail that SpaceX has named Skywalker, moving around carefully and deliberately. The two will not be outside the spacecraft at the same time.

The other two crew members, Scott Poteet, a retired Air Force pilot and the mission’s pilot, and Anna Menon, another SpaceX engineer, will remain in the capsule to manage the umbilical cords that will feed the spacewalkers oxygen and power, and monitor the readings to make sure everything is proceeding properly.

After Mr. Isaacman and Ms. Gillis return inside and close the hatch, the inside of the capsule will be repressurized with oxygen and nitrogen. The entire spacewalk, from letting out all the air to refilling a breathable atmosphere, is expected to take about two hours.

Alban Philips, a maverick self-taught New Zealander, built an analogue computer in 1949, capable of solving 9 simultaneous differential equations in a few minutes and modelling the macroeconomics of a country’s economy - the UK’s first non-human computer.

Now it is being rebuilt

This is a lovely home built automated production line.

Small production lines empowering individuals to produce small run but at scale production runs can make help support maker craft and the small scale idiosyncratic existence of things more possible in a world where competition with offshore manufacture and mass-produced soulless production and craft outcompetes and crushes local craft and garage grown tech.

Bravo !

You miss the similarity.

Modern monopsonmist’s acting like traditional media-barons.

Pretending that heavy editorial control, in this case blaming the algorithm, for shaping the debate, platforming rabble rousing or crime.

Shaping the debate makes them publishers rather than platforms and they should behave responsibly as such not pretend they are akin to ISPS

E.g. heavy handed censorship, Musk is blatant thanks to his egotism it is all about him.

Zuckerberg, Facebook, VK, Telegram hold monopsonies, control is too centralised, too big - consumers and citizens need protection by regulation to protect their human rights and their free speech.

In both cases it is obvious free speech does not mean being allowed to create harm and cause mass panic by absence of responsibility.

I am not a comrade nor subscribe to an ideology using that moniker which I think you imply.

Ideally interoperability and the free market will allow anyone to write and interface to a public API when X and Facebook etc are decentralised.

No interop enshittifies the net for us all.

https://locusmag.com/2021/07/cory-doctorow-tech-monopolies-a...

Musk believes in free speech in the same way that he believes in free Teslas. Free for him, very expensive for everyone else.

In between a stream of tweets on how “Free speech is the bedrock of democracy” and “Censorship is a certainty if Dems win” was another one in which he gloated about a ruling in his libel case against Media Matters for America.

The banning of X in Brazil and the arrest of Telegram boss Pavel Durov won’t stop their lies

This is a mouse study

“Senior author Dr Nathalie Dehorter said the study —

conducted on mice…”

Fascinating work, very promising.

Can you summarise how the model in your paper differs from this implementation of xLSTM ?

https://github.com/huggingface/transformers/issues/27011

This is a very clear and extensive explanation of the hierarchy of difficulty.

Interesting question.

LLMs can be cajoled into producing algorithms.

In fact this is the Chain-of-Thought optimisation.

LLMs give better results when asked for a series of steps to produce a result than when just asked for the result.

To ask if LLMs “think” is an open question and requires a definition of thinking :-)

This a Jupyter notebook.

Well written and comprehensive, N.B. last commit 7 months ago.

Part of the fast.ai free and excellent Deep Learning Course by Jeremy Howard.[0]

This is a pragmatic course with the emphasis on getting things done and no more theory than needed.

Can’t recommend fast.au courses enough - their particular blend of free and pragmatic is unique.

And Jeremy Howard is one of the clearest teachers around. He is engaging, and easy to grok without dumbing down.

[0] https://course.fast.ai/

Bragg & Guests discuss the Great Stink - In Our Time.

https://www.bbc.co.uk/programmes/m001gjcm

Are there any NERF to textured or vertex coloured polygon mesh point cloud tools or papers ?

I.e. can these be assets for traditional game engines ?

Could a sort of photogrammetry like Meshroom or Reality capture do the trick ?

Can you share some of what you have found about the creative process by talking to people with artistic ability ?

What are your ideas about the differences between a human and AI's creative process ?

Are there any similarities, or analagous processes ?

Do you think creators have an kind of latent space where different concepts are inspired by multi-modal inputs ( what sparks inspiration ? e.g. sometimes music or a mood inspires a picture ) and then the creators make different versions of their idea by combining different amounts of different concepts ?

I am not being snarky, I am genuinely interested in views comparing human an AI's creative processes.

Efficiency can be seen as doublespeak for 'those who remain must work harder'.

It could well be that some degree of inefficiency is a good price to pay for all the intangibles that make for a great workplace that produces great work and attracts talented people.

For layoffs to not result in employee disillusionment, actual efficiency must be created - e.g. less people can achieve more with similar effort.

The danger is the new wage bill efficiency is actually creating overwork and mounting technical debt that is unsustainable.

The decision makers are rewarded in the short term for seemingly creating 'efficiency' value when in fact they may have doomed their business by destroying institutional knowledge and lynchpin employees that were not measured on the balance sheet.

The delayed effects are business functionality slowly falters, eventually past tipping points, then cascades to irreversible intermittent faliure.

Shareholder value is maximised in the short term and the policy makers leave well rewarded before the negative implications of the decimation are perceived.

The difficult questions for companies are :

- how to layoff the least ( or negatively ) performant employees instead of a random selection,

- how to not have the best employees leave in the wake of layoffs because the layoffs created a culture of fear and uncertainty,

- how to maintain the motivation of remaining employees so that they remain aligned with company goals, especially in light of the now increased workload for those who remain.

Probably the answer is to share ownership and increase reward with those that remain to increase their engagement.

Currently the majority of companies never do this.

Certainly the remaining employees have to, at best, raise the overwork with management as a problem whereas the management see it as efficiency. At worst the employee will have to fight their corner so as prove their workload is now unsustainable. Without proper support, it can be, that the employee must cut corners to keep up and trust relationships can become adverserial.

One would hope these CEOs have real insight into creating efficiency but in some cases the suspicion is that the real issue is upper tiers who have absolutely no idea how their business actually runs on the shop floor.

The wider societal implications are unemployement and a lack of the (perhaps) vital social function the business provided, and the loss of customer faith in the company to deliver as before.

Orgzly means my Org agenda can give me phone notifications, which means one more thing automated and so one less thing I have to pay attention to

This balance is quite probably a very strong filter on the Drake equation.

Climate and Biosphere is delicate with lots of tipping points and cascades to uninhabitable.

A few billion years ago, the inner planets of the Sol system were capable of supporting abundant surface life - now only Terra.