All Courses - Page 75 of 395 - Analytics Campus

West Coast Stat Views (on Observational Epidemiology and extra): The Voss Chronicles

Epidemiology

Dr. Mike

February 11, 2026

West Coast Stat Views (on Observational Epidemiology and extra): The Voss Chronicles

I used to be watching some movies to unwind earlier than bedtime final evening, and the next advert popped up. I usually would have skipped it, however one thing about it piqued my curiosity and, because it was lower than 3 minutes, I made a decision to let it play.

I used to be not upset.

The easy, man-of-the-soil pitchbot defined that he had been working for an inexplicably rich farmer who claimed that the key of his success was a system that had allowed him to win the lottery a number of occasions utilizing AI. The implausibility of the pitch would have been sufficient to carry my curiosity, nevertheless it obtained higher. The genius behind this wonderful AI was a scientist named Dr. Leonard Voss.

Readers of the weblog and college students of generative AI esoterica will keep in mind Elara Voss:

What’s so odd about that is that–for a reputation now so frequent throughout the
megaplatforms–before 2023, “Elara Voss” didn’t exist. There isn’t a
individual named Elara Voss in america. No delivery certificates has
ever been issued below that identify; in the event you seek for it in public information
databases, you’ll flip up no outcomes. There aren’t even any characters
named “Elara Voss” in any e book printed earlier than 2023. Till two years
in the past, the 2 phrases didn’t ever seem subsequent to one another even by
accident.

However in the event you direct nearly any L.L.M. to generate a
sci-fi story or narrative for you, it’s going to identify the principle character
“Elara Voss”–or an identical variation like “Elara Vex,” “Elena Voss,” or
“Elias Vance”–with an alarming diploma of frequency.

I am unable to say for sure that Dr. Leonard Voss is an ancestor of Elena’s, however a fast Google search did present him popping up in comparable roles. Maybe some future LLM-generated novel will reveal that the Voss household dynasty was constructed on Powerball winnings.

Programming an estimation command in Stata: Dealing with issue variables in optimize()

Econometrics

Dr. Mike

February 11, 2026

Programming an estimation command in Stata: Dealing with issue variables in optimize()

(
newcommand{xb}{{bf x}}
newcommand{betab}{boldsymbol{beta}})I focus on a technique for dealing with issue variables when performing nonlinear optimization utilizing optimize(). After illustrating the difficulty brought on by issue variables, I current a technique and apply it to an instance utilizing optimize().

That is the twenty publish within the sequence Programming an estimation command in Stata. I like to recommend that you simply begin at first. See Programming an estimation command in Stata: A map to posted entries for a map to all of the posts on this sequence.

How poisson handles issue variables

Contemplate the Poisson regression by which I embrace a full set of indicator variables created from the specific variable youngsters and a continuing time period.

Instance 1: Collinear issue variables


. clear all

. use accident3

. poisson accidents cvalue ibn.youngsters site visitors, coeflegend
word: 3.youngsters omitted due to collinearity

Iteration 0:   log probability = -546.35782
Iteration 1:   log probability = -545.11016
Iteration 2:   log probability = -545.10898
Iteration 3:   log probability = -545.10898

Poisson regression                              Variety of obs     =        505
                                                LR chi2(5)        =     361.62
                                                Prob > chi2       =     0.0000
Log probability = -545.10898                     Pseudo R2         =     0.2491

------------------------------------------------------------------------------
   accidents |      Coef.  Legend
-------------+----------------------------------------------------------------
      cvalue |  -.6582924  _b[cvalue]
             |
        youngsters |
          0  |   3.233932  _b[0bn.kids]
          1  |   1.571582  _b[1.kids]
          2  |   1.659241  _b[2.kids]
          3  |          0  _b[3o.kids]
             |
     site visitors |   .1383977  _b[traffic]
       _cons |  -2.518175  _b[_cons]
------------------------------------------------------------------------------

The total set of indicator variables is collinear with the fixed time period. The output exhibits that no variables have been dropped however the title 3o.youngsters specifies that 3.youngsters was omitted. Omitted variables aren’t dropped; as an alternative their coefficients are constrained to zero.

Specifying variables as omitted as an alternative of dropping them permits postestimation options equivalent to margins to work correctly.

For the case in instance 1, poisson is maximizing the log-likelihood operate topic to constraint that (beta_{3.youngsters}=0). When it comes to the parameter vector (betab), I signify this constraint by

[
left[begin{matrix}
0&0&0&0&1&0&0
end{matrix}right]
betab’ = 0
]

the place (betab=(
betab_{cvalue},
betab_{0.youngsters},
betab_{1.youngsters},
betab_{2.youngsters},
betab_{3.youngsters},
betab_{site visitors},
betab_{_cons}))

Usually, I can signify (q) linear equality constraints on a (1times okay) parameter vector as

[
{bf C}betab’ = {bf c}
]

the place ({bf C}) is a (qtimes okay) matrix and ({bf c}) is a (qtimes 1) vector. These constraints are conveniently represented as (widetilde{bf C}=left[{bf C},{bf c}right]).

I now present tips on how to use optimize() to unravel optimization issues topic to linear equality constraints by placing (widetilde{bf C}) into the optimize object. In code block 1, I take advantage of optimize() to maximise the Poisson log-likelihood operate for the issue in instance 1. Code block 1 augments instance 3 in Programming an estimation command in Stata: Utilizing optimize() to estimate Poisson parameters through the use of optimize_init_constraints() to impose a linear equality constraint on the coefficient vector.

Code block 1: Linear equality constraints in optimize()


mata:
void plleval3(actual scalar todo, actual vector b,     ///
              actual vector y,    actual matrix X,     ///
              val, grad, hess)
{
    actual vector  xb

    xb = X*b'
   val = -exp(xb) + y:*xb - lnfactorial(y)
}

y  = st_data(., "accidents")
X  = st_data(., "cvalue ibn.youngsters site visitors")
X  = X,J(rows(X), 1, 1)

C  = e(5, 7)
c  = 0
Ct = C,c

S  = optimize_init()
optimize_init_argument(S, 1, y)
optimize_init_argument(S, 2, X)
optimize_init_evaluator(S, &plleval3())
optimize_init_evaluatortype(S, "gf0")
optimize_init_params(S, J(1, 7, .01))
optimize_init_constraints(S, Ct)

bh = optimize(S)
optimize_result_params(S)
finish

Solely line 13, strains 16–18, and line 26 differ from the code in instance 3 in Programming an estimation command in Stata: Utilizing optimize() to estimate Poisson parameters. Line 13 illustrates that st_data() can create the indicator variables from the issue variable ibn.youngsters. Traces 16–18 outline the constraint matrix (widetilde{bf C}) for this downside. Line 26 places (widetilde{bf C}) into the optimize() object S, which causes optimize() to maximise the Poisson log-likelihood operate with evaluator plleval3() topic to constraints laid out in matrix Ct.

Instance 2 illustrates that code block 1 reproduces the purpose estimates reported in instance 1.

Instance 2: Linear equality constraints in optimize()


. do laptop

. mata:
------------------------------------------------- mata (kind finish to exit) -----
: void plleval3(actual scalar todo, actual vector b,     ///
>               actual vector y,    actual matrix X,     ///
>               val, grad, hess)
> {
>     actual vector  xb
>
>     xb = X*b'
>    val = -exp(xb) + y:*xb - lnfactorial(y)
> }
word: argument todo unused
word: argument grad unused
word: argument hess unused

:
: y  = st_data(., "accidents")

: X  = st_data(., "cvalue ibn.youngsters site visitors")

: X  = X,J(rows(X), 1, 1)

:
: C  = e(5, 7)

: c  = 0

: Ct = C,c

:
: S  = optimize_init()

: optimize_init_argument(S, 1, y)

: optimize_init_argument(S, 2, X)

: optimize_init_evaluator(S, &plleval3())

: optimize_init_evaluatortype(S, "gf0")

: optimize_init_params(S, J(1, 7, .01))

: optimize_init_constraints(S, Ct)

:
: bh = optimize(S)
Iteration 0:   f(p) = -845.47138
Iteration 1:   f(p) = -572.68676
Iteration 2:   f(p) = -545.68381
Iteration 3:   f(p) = -545.11241
Iteration 4:   f(p) = -545.10898
Iteration 5:   f(p) = -545.10898
: optimize_result_params(S)
                  1              2              3              4
    +-------------------------------------------------------------
  1 |  -.6582923624    3.233932519    1.571581623     1.65924145
    +-------------------------------------------------------------
                  5              6              7
     ----------------------------------------------+
  1               0      .13839766   -2.518174926  |
     ----------------------------------------------+

: finish
-------------------------------------------------------------------------------

.
finish of do-file

Code block 1 exhibits tips on how to use a linear equality constraint to deal with collinear variables after we know which variables are omitted. Within the code for an estimation command, we should

discover which variables will probably be omitted, and
create the constraint matrix (widetilde{bf C}) that imposes the constraints implied by omitting these variables.

Instance 3 illustrates that _rmcoll shops a listing of variables that identifies which variables will probably be omitted in r(varlist), thereby fixing downside 1.

Instance 3: Utilizing _rmcoll to establish omitted variables


. _rmcoll cvalue ibn.youngsters site visitors, develop
word: 3.youngsters omitted due to collinearity

. return listing

scalars:
          r(k_omitted) =  1

macros:
            r(varlist) : "cvalue 0bn.youngsters 1.youngsters 2.youngsters 3o.youngsters site visitors"

. native cnames "`r(varlist)' _cons"

I specified the choice develop in order that _rmcoll would develop any issue variables. The expanded variable listing within the native r(varlist) identifies 3.youngsters as a variable that should be omitted. I then put this expanded variable listing, augmented by the title _cons, within the native macro cnames.

Right here is a top level view for the answer to downside 2 that I current in examples 4–6.

In instance 4, I create the Stata vector bt, whose column names are contained in cnames.
In instance 5, I take advantage of _ms_omit_info to create the Stata vector bto, which signifies which variables will probably be omitted from bt.
In instance 6, I create a Mata matrix specifying the constraints from bto.

Now for the small print, starting with instance 4.

Instance 4: Placing the coefficient names on a Stata vector


. matrix bt = J(1, 7, 0)

. matrix colnames bt = `cnames'

. matrix listing bt

bt[1,7]
                   0.       1.       2.      3o.
     cvalue     youngsters     youngsters     youngsters     youngsters  site visitors    _cons
r1        0        0        0        0        0        0        0

cnames accommodates the names of the coefficients for this downside, so I create a conformable row vector bt, make cnames the column names on bt, and show bt. The values in bt don’t matter; the column names are the essential data.

In instance 5, _ms_omit_info makes use of the column names on a Stata vector to create the vector r(omit), which specifies which variables are omitted.

Instance 5: Making a vector that signifies omitted variables


. matrix bt = J(1, 8, 0)

. matrix colnames bt = `cnames'

. _ms_omit_info bt

. return listing

scalars:
             r(k_omit) =  1

matrices:
               r(omit) :  1 x 8

. matrix bto = r(omit)

. matrix listing bto

bto[1,8]
    c1  c2  c3  c4  c5  c6  c7  c8
r1   0   0   0   0   1   0   0   0

A component of r(omit) is 1 if the corresponding variable is omitted. A component of r(omit) is 0 if the corresponding variable will not be omitted. I put a duplicate of r(omit) in bto.

In instance 6, I create a constraint matrix from bto. The loop in instance 6 will create the constraint matrix implied by any r(omit) vector created by _ms_omit_info.

Instance 6: Making a constraint matrix from r(omit)


. mata:
------------------------------------------------- mata (kind finish to exit) -----
: mo = st_matrix("bto")

: ko = sum(mo)

: p  = cols(mo)

: if (ko>0) {
>     Cm   = J(0, p, .)
>     for(j=1; j<=p; j++) {
>         if (mo[j]==1) {
>             Cm  = Cm  e(j, p)
>         }
>     }
>     Cm = Cm, J(ko, 1, 0)
> }
> else {
>     Cm = J(0,p+1,.)
> }

: "Constraint matrix is "
  Constraint matrix is 

: Cm
       1   2   3   4   5   6   7   8   9
    +-------------------------------------+
  1 |  0   0   0   0   1   0   0   0   0  |
    +-------------------------------------+

: finish
-------------------------------------------------------------------------------

After copying bto to the Mata vector mo, I put the variety of constraints within the scalar ko and the variety of parameters in p. If there are constraints, I initialize Cm to be a matrix with zero rows and p columns, use a for loop to iteratively append a brand new row corresponding to every constraint recognized in mo, and end by appending a ko (instances) 1 column of zeros on to Cm. If there are not any constraints, I put a matrix with zero rows and p+1 columns in Cm.

No matter whether or not there are any omitted variables, I can put the Cm matrix created by the tactic in instance 6 into an optimize() object. If there are not any omitted variables, Cm may have zero rows, and no constraints will probably be imposed. If there are omitted variables, Cm may have ko rows, and the constraints for the omitted variables will probably be imposed.

Code block 2 combines these items right into a coherent instance.

Code block 2: Placing all of it collectively


clear all
use accident3
native depvar    "accidents"
native indepvars "cvalue ibn.youngsters site visitors"
_rmcoll `indepvars', develop
native cnames "`r(varlist)' _cons"
native p   : phrase depend `cnames'
matrix bt = J(1, `p', 0)
matrix colnames bt = `cnames'
_ms_omit_info bt
matrix bto = r(omit)

mata:
void plleval3(actual scalar todo, actual vector b,     ///
              actual vector y,    actual matrix X,     ///
              val, grad, hess)
{
    actual vector  xb

    xb = X*b'
   val = -exp(xb) + y:*xb - lnfactorial(y)
}

y  = st_data(., "`depvar'")
X  = st_data(., "`indepvars'")
X  = X,J(rows(X), 1, 1)

mo = st_matrix("bto")
ko = sum(mo)
p  = cols(mo)
if (ko>0) {
    Ct   = J(0, p, .)
    for(j=1; j<=p; j++) {
        if (mo[j]==1) {
            Ct  = Ct  e(j, p)
        }
    }
    Ct = Ct, J(ko, 1, 0)
}
else {
    Ct = J(0,p+1,.)
}

S  = optimize_init()
optimize_init_argument(S, 1, y)
optimize_init_argument(S, 2, X)
optimize_init_evaluator(S, &plleval3())
optimize_init_evaluatortype(S, "gf0")
optimize_init_params(S, J(1, 7, .01))
optimize_init_constraints(S, Ct)

bh = optimize(S)
optimize_result_params(S)
finish

Traces 1–2 drop all of the objects that I created in earlier examples and browse the accident3 dataset into reminiscence. Traces 3–4 create locals to carry the dependent variable and the impartial variables.

Traces 5–6 use _rmcoll to establish which variables must be omitted and put the listing of names within the native macro cnames, as in instance 3. Traces 7–9 create bt, whose column names specify which variables must be omitted, as in instance 4. Traces 10–11 create bto, whose entries specify which variables must be omitted, as in instance 5. Traces
28–42 create the constraint matrix Ct from bto, as in instance 6.

Line 50 places Ct into the optimize() object.

Instance 7 illustrates that the code in code block 2 reproduces the
beforehand obtained outcomes.

Instance 7:Placing all of it collectively


. do pc2

. clear all

. use accident3

. native depvar    "accidents"

. native indepvars "cvalue ibn.youngsters site visitors"

. _rmcoll `indepvars', develop
word: 3.youngsters omitted due to collinearity

. native cnames "`r(varlist)' _cons"

. native p   : phrase depend `cnames'

. matrix bt = J(1, `p', 0)

. matrix colnames bt = `cnames'

. _ms_omit_info bt

. matrix bto = r(omit)

.
. mata:
------------------------------------------------- mata (kind finish to exit) -----
: void plleval3(actual scalar todo, actual vector b,     ///
>               actual vector y,    actual matrix X,     ///
>               val, grad, hess)
> {
>     actual vector  xb
>
>     xb = X*b'
>    val = -exp(xb) + y:*xb - lnfactorial(y)
> }
word: argument todo unused
word: argument grad unused
word: argument hess unused

:
: y  = st_data(., "`depvar'")

: X  = st_data(., "`indepvars'")

: X  = X,J(rows(X), 1, 1)

:
: mo = st_matrix("bto")

: ko = sum(mo)

: p  = cols(mo)

: if (ko>0) {
>     Ct   = J(0, p, .)
>     for(j=1; j         if (mo[j]==1) {
>             Ct  = Ct  e(j, p)
>         }
>     }
>     Ct = Ct, J(ko, 1, 0)
> }
> else {
>     Ct = J(0,p+1,.)
> }

:
: S  = optimize_init()

: optimize_init_argument(S, 1, y)

: optimize_init_argument(S, 2, X)

: optimize_init_evaluator(S, &plleval3())

: optimize_init_evaluatortype(S, "gf0")

: optimize_init_params(S, J(1, 7, .01))

: optimize_init_constraints(S, Ct)

:
: bh = optimize(S)
Iteration 0:   f(p) = -845.47138
Iteration 1:   f(p) = -572.68676
Iteration 2:   f(p) = -545.68381
Iteration 3:   f(p) = -545.11241
Iteration 4:   f(p) = -545.10898
Iteration 5:   f(p) = -545.10898

: optimize_result_params(S)
                  1              2              3              4
    +-------------------------------------------------------------
  1 |  -.6582923624    3.233932519    1.571581623     1.65924145
    +-------------------------------------------------------------
                  5              6              7
     ----------------------------------------------+
  1               0      .13839766   -2.518174926  |
     ----------------------------------------------+

: finish
-------------------------------------------------------------------------------

.
finish of do-file

Achieved and undone

I mentioned a technique for dealing with issue variables when performing nonlinear optimization utilizing optimize(). In my subsequent publish, I implement these strategies in an estimation command for Poisson regression.

The demise of reactive IT: How predictive engineering will redefine cloud efficiency in 10 years

Dr. Mike

February 11, 2026

The demise of reactive IT: How predictive engineering will redefine cloud efficiency in 10 years

CLOSED-LOOP FEEDBACK SYSTEM

This pipeline captures how knowledge is ingested, modeled, predicted and acted upon in a real-time system.

Reactive vs predictive lifecycle

Reactive IT:

Occasion Happens → Alert → People Reply → Repair → Postmortem

Predictive IT:

Predict → Stop → Execute → Validate → Study

Predictive Kubernetes workflow

Metrics + Traces + Occasions

│

▼

Forecasting Engine

(Math-driven future projection)

│

▼

Causal Reasoning Layer

(Dependency-aware impression evaluation)

│

▼

Prediction Engine Output

“Node Pool X will saturate in 25 minutes”

│

▼

Autonomous Remediation Actions

Pre-scaling nodes
Pod rebalancing
Cache priming
Site visitors shaping

│

▼

Validation

The longer term: Autonomous infrastructure and zero-war-room operations

Predictive engineering will usher in a brand new operational period the place outages develop into statistical anomalies relatively than weekly realities. Methods will not look forward to degradation, they may preempt it. Conflict rooms will disappear, changed by steady optimization loops. Cloud platforms will behave like self-regulating ecosystems, balancing sources, site visitors and workloads with anticipatory intelligence.

In SAP environments, predictive fashions will anticipate period-end compute calls for and autonomously modify storage and reminiscence provisioning. In Kubernetes, predictive scheduling will stop node imbalance earlier than it types. In distributed networks, routing will adapt in actual time to keep away from predicted congestion. Databases will modify indexing methods earlier than question slowdowns accumulate.

The long-term trajectory is unmistakable: autonomous cloud operations.

Predictive engineering just isn’t merely the subsequent chapter in observability, it’s the basis of absolutely self-healing, self-optimizing digital infrastructure.

Organizations that undertake this mannequin early will get pleasure from a aggressive benefit measured not in small increments however in orders of magnitude. The way forward for IT belongs to methods that anticipate, not methods that react.

This text is revealed as a part of the Foundry Professional Contributor Community.
Wish to be a part of?

The Obtain: Contained in the QuitGPT motion, and EVs in Africa

Artificial Intelligence

Dr. Mike

February 11, 2026

The Obtain: Contained in the QuitGPT motion, and EVs in Africa

The recognition of business nuclear reactors has surged in recent times as worries about local weather change and power independence drowned out issues about meltdowns and radioactive waste.

The issue is, constructing nuclear energy crops is pricey and sluggish.

A brand new technology of nuclear energy know-how might reinvent what a reactor seems like—and the way it works. Advocates hope that new tech can refresh the trade and assist change fossil fuels with out emitting greenhouse gases.

That is our newest story to be become a MIT Know-how Overview Narrated podcast, which we’re publishing every week on Spotify and Apple Podcasts. Simply navigate to MIT Know-how Overview Narrated on both platform, and observe us to get all our new content material because it’s launched.

The must-reads

I’ve combed the web to search out you at this time’s most enjoyable/vital/scary/fascinating tales about know-how.

1 Social media giants have agreed to be rated on teen security
Meta, TikTok and Snap will endure unbiased assessments over how successfully they defend the psychological well being of adlescent customers. (WP $)
+ Discord, YouTube, Pinterest, Roblox and Twitch have additionally agreed to be graded. (LA Instances $)

2 The FDA has refused to evaluation Moderna’s mRNA flu vaccine
It’s the newest in an extended line of anti-vaccination strikes the company is making. (Ars Technica)
+ Consultants fear it’ll have a knock-on impact on funding in future vaccines. (The Guardian)
+ Moderna says it was blindsided by the choice. (CNN)

3 EV battery factories are pivoting to manufacturing power cells
Power storage techniques are in, electrical automobiles are out. (FT $)

4 Why OpenAI killed off ChatGPT’s 4o mannequin
The qualities that make it engaging for some customers make it extremely dangerous for others. (WSJ $)
+ Bereft customers have arrange their very own Reddit group to mourn. (Futurism)
+ Why GPT-4o’s sudden shutdown left folks grieving. (MIT Know-how Overview)

5 Drug cartels have began laundering cash by means of crypto
And legislation enforcement is struggling to cease them. (Bloomberg $)

6 Morocco desires to construct an AI for Africa
The nation’s Minister of Digital Transition has a plan. (Remainder of World)
+ What Africa must do to turn into a significant AI participant. (MIT Know-how Overview)

AI economic system: How Claude Code may upend white-collar work in 2026

Technology

Dr. Mike
-

February 11, 2026
0

It’s February 2020 once more.

An exponential course of is in movement — one that can inevitably shake the world to its core — and upend our economic system, politics, and social lives. But most individuals are nonetheless going about their enterprise, oblivious as dinosaurs to a descending asteroid.

That is what many in and round the AI trade imagine, anyway.

Besides, on this telling, the invisible drive that’s about to vary our world isn’t a virus that can rip by the inhabitants after which ebb. Fairly, it’s an data know-how that can irreversibly remodel (if not extinguish) white-collar labor, speed up scientific progress, destabilize political techniques, and, maybe, get us all killed.

After all, such apocalyptic chatter has all the time hummed within the background of the AI discourse. However it’s grown a lot louder in current weeks.

• AI “brokers” like Claude Code can autonomously full complicated initiatives — not simply reply questions — making them potential substitutes for expert staff.
• Buyers are actually treating agentic AI as an existential menace to many incumbent software program and consulting companies.
• If AI’s capabilities preserve enhancing at an exponential fee, issues may get actually bizarre by 2027.

SemiAnalysis, a distinguished chip trade commerce publication, declared final Thursday that AI progress had hit an “inflection level.” At Cisco Programs’ AI summit that very same week, OpenAI CEO Sam Altman declared, “that is the primary time I felt one other ChatGPT second — a transparent glimpse into the way forward for data work.” Not lengthy earlier than these remarks, Altman’s rival, Anthropic CEO Dario Amodei, wrote that current breakthroughs had made it clear that we’re solely “a couple of years” away from the purpose when “AI is healthier than people at basically every thing.” (Disclosure: Vox Media is one among a number of publishers which have signed partnership agreements with OpenAI. Our reporting stays editorially impartial. The Vox part Future Excellent is funded partly by the BEMC Basis, whose main funder was additionally an early investor in Anthropic; they don’t have any editorial enter into our content material.)

In a succinct abstract of the tech-savvy’s new zeitgeist, the efficient altruist author Andy Masley posted on X, “I do know everybody’s saying it’s feeling lots like February 2020 however it’s feeling lots like February 2020.”

Critically, tech pundits and executives aren’t alone in considering that one thing simply modified. In current weeks, software program companies noticed their inventory costs plunge, as merchants determined that AI would quickly render a lot of them out of date.

Not way back, the standard knowledge round AI’s near-term results sounded radically totally different. For a lot of final 12 months, trade analysts and journalists warned that AI had turn into a bubble ripe for popping.

In any case, main labs’ capital expenditures had been far outpacing their earnings; OpenAI alone was slated to take a position $1.4 trillion in infrastructure over the following eight years, even because it collected solely $20 billion in annual recurring income. These gargantuan investments would solely repay if demand for AI providers skyrocketed.

And the know-how’s business potential appeared unsure. At the same time as enterprise capitalists waxed rhapsodic about AI’s transformative powers, official financial information confirmed its impacts on productiveness and employment had been marginal, at finest.

So, what modified? Why accomplish that many traders, entrepreneurs, and analysts — together with some who’d subscribed to the “AI bubble” thesis mere months in the past — now imagine that synthetic intelligence resides as much as its hype?

The reply, in three phrases, is the “agentic” revolution.

AI brokers, briefly defined

Till just lately, public-facing AI techniques had been basically passive. You typed a query to ChatGPT and the robotic replied, then awaited your subsequent instruction. The expertise was a bit like texting with an infinitely huge and sycophantic encyclopedia — one that might streamline your presentation, repair your code, diagnose your rash, or validate your perception {that a} malevolent cabal had implanted a digicam in your mom’s printer.

These chatbots had actual financial utility. However additionally they had strict limitations. Gemini may draft your e mail, nevertheless it couldn’t ship it. Claude may generate code, nevertheless it couldn’t run it, see what broke, revise this system, after which give it one other shot.

In different phrases, the chatbots may automate duties however not complicated, time-intensive initiatives. To finish the latter, they wanted a human to carry their figurative palms and challenge directions at every step within the course of.

Then, final 12 months, commercially viable AI brokers hit the market.

These new techniques are extra autonomous and dynamic than their predecessors. Fairly than answering one discrete immediate after which awaiting additional orders, Claude Code or OpenAI’s Codex receives a broad goal — reminiscent of “detect and repair the bug that’s crashing our app” or “monitor regulatory filings and flag something related to our enterprise” or “make a 3D flying sport” — after which figures out how one can obtain its mission.

Put in another way, these AIs operate much less like souped-up engines like google and extra like junior staffers. They’ll independently resolve which steps to take subsequent, make the most of instruments (like code editors, spreadsheets, or firm databases), check whether or not their plan labored, attempt one other strategy if it fails, and proceed iterating till their job is finished.

Why agentic AI is a gamechanger

That is what the large labs had lengthy promised however did not ship: Machines that might not solely complement high-skilled staff however — a minimum of in some instances — dramatically outperform them.

Over the course of 2025, AI brokers solely grew extra succesful. By 12 months’s finish, consciousness of the instruments’ energy had damaged containment: Influencers with no engineering expertise realized they might “vibe code” complete web sites, apps, and video games.

This month, CNBC supplied a very vivid illustration of the brand new techniques’ transformative potential. Two of the outlet’s journalists — every with none coding expertise — got down to construct a competitor to Monday.com, a challenge administration platform then valued at $5 billion. They advised Claude Code to analysis Monday, establish its major options, and recreate them. Inside an hour, that they had constructed a useful alternative for the agency’s software program. Since CNBC’s story revealed final week, Monday’s inventory value has fallen by roughly 20 %.

So, that is one cause why many technologists and commentators are predicting large, near-term AI-induced disruption: Even when AI progress stopped at the moment, the adoption of current techniques would abruptly devalue many companies and white-collar staff.

As SemiAnalysis put the latter level:

One developer with Claude Code can now do what took a crew a month.

The price of Claude Professional or ChatGPT is $20 {dollars} a month, whereas a Max subscription is $200 {dollars} respectively. The median US data employee prices ~350-500 {dollars} a day totally loaded. An agent that handles even a fraction of their workflow a day at ~6-7 {dollars} is a 10-30x ROI not together with enchancment in intelligence.

What’s extra, as Monday.com just lately found, it isn’t simply the data economic system’s staff who’re liable to displacement. At first, traders had largely assumed that AI brokers would profit incumbent software program corporations and consulting companies by growing their productiveness: They’d now be capable of roll out extra apps and audits with fewer staff.

However in current weeks, many merchants realized that agentic AI may simply as simply render such companies irrelevant: Why pay Gartner for a analysis report — or Asana for work administration software program — when Claude Code can present you each at a fraction of the fee? Such reasoning has led to selloffs in software program and consulting shares, with Gartner and Asana every shedding greater than one-third of their worth over the previous month.

On the similar time, AI brokers have eased Wall Avenue’s fears of an artificial-intelligence bubble: The concept that demand is poised to soar for Claude, ChatGPT, and Gemini — and the info facilities that help them — appears much less far-fetched than it did six months in the past.

If we automate automation, issues will begin to get bizarre

Nonetheless, the first driver of Silicon Valley’s millenarian rhetoric isn’t agentic AI’s current capacities, however reasonably, its potential future skills.

No corporations are embracing AI brokers extra vigorously than the highest labs themselves. Engineers at Anthropic and OpenAi have stated that almost 100% of their code is now AI-generated.

To some, this implies that AI progress received’t proceed in a gentle march a lot as a series response: As AI brokers construct their very own successors, every advance will speed up the following, triggering a self-reinforcing suggestions loop by which innovation compounds on itself.

By some measures, AI’s capacities are already rising exponentially. METR, a nonprofit artificial-intelligence analysis group, gauges AI efficiency by measuring the size of coding duties that fashions can full with 50 % success. It finds that this size has been doubling each 7 months.

The human thoughts struggles to internalize the implications of exponential change. At the beginning of March 2020, Covid instances had been doubling each two to 3 days within the US. But absolutely the variety of instances remained tiny at the beginning of the month; on March 1, there have been solely about 40 confirmed instances in the entire nation. Many People had been subsequently caught unaware when, by April 1, greater than 200,000 of their compatriots had been struck ailing by the virus.

These bullish on AI progress imagine People are as soon as once more sleeping on the pace and scale of what’s to return. On this view, as spectacular as AI brokers’ present capabilities are, they’ll pale compared to these on the fingertips of everybody with an web connection this December. As with the pandemic, the total penalties of an on the spot industrial revolution are sure to be each immense and unforeseeable.

The robotic apocalypse (and/or utopia) isn’t essentially nigh

There’s little query that agentic AI goes to reshape the white-collar economic system. Whether or not it has introduced us to the cusp of a courageous new world, nonetheless, is much less sure.

There are lots of causes to assume that AI’s near-term impacts might be smaller and slower than Silicon Valley’s bulls (and catastrophists) now imagine.

First, AI nonetheless makes errors. And this fallibility arguably constrains its potential for changing human staff within the right here and now. An autonomous agent would possibly be capable of execute the suitable commerce, ship the specified e mail, and exchange the errant line of code 9 instances out of 10. If that different time it stakes all of your agency’s capital on Dogecoin, tells off your prime consumer, and introduces a safety vulnerability into your app, nonetheless, you’re most likely gonna retain numerous human supervision over your highest-stakes initiatives.

Second, institutional inertia tends to sluggish adoption of latest applied sciences. Though turbines grew to become widespread within the late nineteenth century, it took many years for factories to reorganize round electrical energy. Equally, whereas tech companies could have little hassle integrating agentic AI into their workflows, legacy firms could take longer to regulate. And in some key sectors — reminiscent of well being care and legislation — laws could additional constrain AI deployment.

Most critically, it’s not clear whether or not AI’s capabilities will proceed rising exponentially. Loads of previous applied sciences loved compounding returns for some time, solely to plateau.

However, the bulls’ case has gotten stronger. Right this moment’s AI techniques are already highly effective sufficient to remodel many industries. And tomorrow’s will certainly be much more succesful. If celebrations of the singularity are untimely, preparations for one thing prefer it are actually overdue.

May the stays of a ‘useless’ comet nonetheless be within the photo voltaic system? Astronomers are nonetheless looking out 6 years later

Science

Dr. Mike
-

February 11, 2026
0

The destiny of a comet that was predicted to cross near Earth stays a thriller 5 years after its dramatic breakup within the internal photo voltaic system — however some astronomers assume part of it would nonetheless be on the market.

In early 2020, astronomers found the icy traveler, often known as C/2019 Y4 ATLAS, and predicted it would present a night-sky spectacle that will enliven everybody’s COVID-19 pandemic lockdown: a comet seen with the unaided eye because it handed inside 23 million miles (37.5 million kilometers) of the solar, or about one quarter of the gap at which Earth orbits our star. However then the comet broke into dozens of items, leaving would-be observers hanging — and leaving astronomers questioning whether or not there may nonetheless be something substantial left of our ill-fated icy customer.

Chances are you’ll like

Lowell Discovery Telescope (a 4.3-meter telescope in Arizona) and nightly pictures from the Zwicky Transient Facility (which makes a wide-view scan of the northern sky each two nights, searching for altering or short-lived objects like comets and supernovas) turned up nothing. However that does not imply there’s nothing left of C/2019 Y4; it would simply imply that what’s left is smaller than the smallest fragment these telescopes would have been capable of see, which comes out to about half a kilometer extensive.

Moreover fixing an intriguing astronomical thriller, this new research of C/2019 Y4 provides some clues about what occurs when comets break aside within the intense warmth close to the solar, in addition to an opportunity to review the millennia-long decline of an historic comet household (C/2019 Y4 may be a fraction of a bigger comet which broke up 1000’s of years in the past, in response to a 2021 research).

“The unsure destiny of C/2019 Y4 raises an intriguing query,” the astronomers wrote within the research. “What number of presumably disrupted comets have actually utterly disrupted, and will any of them have really survived with a lowered, inactive nucleus?”

Within the case of C/2019 Y4, the reply to that second query could also be sure: it is potential {that a} fragment of comet, lower than half a kilometer extensive, may nonetheless be tracing its bigger dad or mum’s lengthy path round the solar.

Breaking area information, the most recent updates on rocket launches, skywatching occasions and extra!

began getting brighter in a short time in early 2020 because it flew towards the internal photo voltaic system, and astronomers excitedly predicted that it might be seen with the unaided eye by the point it made its closest cross to Earth in late Might.

After which, like the remainder of us, C/2019 Y4 all of the sudden fell aside in late April 2020.

The Digital Telescope Venture captured this view of Comet Atlas’ shattered nucleus on April 11, 2020. (Picture credit score: Gianluca Masi/Digital Telescope Venture (www.virtualtelescope.eu))

Within the aftermath, astronomers used the Hubble House Telescope and different observatories all over the world to trace a few dozen items of the shattered comet, in what gave the impression to be grouped in 4 primary clusters of icy particles. However a kind of clusters later turned out to be a glitch within the knowledge, and one other solely lasted for just a few days earlier than utterly dissipating. That left two extra particles clusters, dubbed fragment A and fragment B.

Chances are you’ll like

A Hubble House Telescope pictures of comet C/2019 Y4 (ATLAS), taken on April 20, 2020, displaying the strong nucleus of the comet breaking up. (Picture credit score: NASA, ESA, Quanzhi Ye (UMD); Picture Processing: Alyssa Pagan (STScI))

Astronomers’ final glimpse of any piece of C/2019 Y4’s icy particles was on June 8, 2020, in pictures from NASA’s STEREO spacecraft, 9 days after the comet’s closest method to the solar. On the time, the comet’s nucleus undoubtedly regarded “utterly disrupted,” Cordova Quijano and colleagues wrote. The lingering query is what occurred to the nucleus after these observations.

By now, Fragment A might be nothing however a slowly-spreading cloud of gasoline and perhaps some mud grains. Inside the first three days after the breakup, the chunks of former comet nucleus that made up fragment A appeared to lose about 70% of their mass (as a result of, once more, ice sublimates, and smaller chunks are likely to sublimate sooner than large ones).

Simply earlier than perihelion, in late Might 2020, the largest chunk of fragment B was about 0.75 miles, or 1.2 kilometers extensive. By the point of Cordova Quijano and colleagues’ observations in late August and mid-October 2020, it was clear that “fragment B had undergone additional main disintegration,” however it’s nonetheless not clear precisely how a lot. Cordova Quijano and co-authors could not spot any hint of fragment B of their Lowell or Zwicky knowledge, which may imply there’s nothing left — or that the remaining fragment is lower than half a kilometer extensive.

“We can’t conclude from the out there knowledge whether or not any sizable fragments nonetheless exist,” they wrote. “Noticed disintegration occasions have produced long-lasting fragments as small as 0.3 kilometers diameter, which is smaller than our detection restrict.”

catch the following one

For astronomers, C/2019 Y4’s dramatic breakup supplied a uncommon probability to observe a comet disintegrate. Thus far, they’ve solely gotten to look at this dramatic phenomenon a handful of occasions: three confirmed and 4 solely suspected. Of these 4, astronomers haven’t any actual concept what occurred after the breakup — like whether or not any giant chunks survived lengthy sufficient to make it out of the new internal photo voltaic system — and in response to Cordova Quijano and his colleagues, that is principally due to a scarcity of follow-up observations to verify the comets’ fates.

The researchers wrote that about two or three months after every comet handed “behind” the solar from our viewpoint, then emerged once more, it ought to have been simpler for telescopes to see. This might have been an ideal time to search for surviving fragments — or their absence. Such observations would have confirmed the comets’ demise and likewise make clear whether or not smaller items of their shattered nuclei would maintain orbiting the solar as mini-comets.

“For C/2019 Y4, a deep search proper after the photo voltaic conjunction (comparable to instantly following the preliminary shallower search in early August of 2020) may have conclusively decided the state of the remnant,” they wrote of their latest paper. “Equally, devoted deep searches could be useful in closing circumstances like the opposite three comets and would offer insights into comet disruption dynamics.”

It is somewhat too late to do this for C/20129 Y4, however the research provides a heads-up to astronomers to be ready for these sorts of observations the following time a comet falls aside on its approach by way of the internal photo voltaic system.

Examine: Platforms that rank the newest LLMs could be unreliable | MIT Information

Machine Learning

Dr. Mike
-

February 11, 2026
0

A agency that desires to make use of a big language mannequin (LLM) to summarize gross sales studies or triage buyer inquiries can select between lots of of distinctive LLMs with dozens of mannequin variations, every with barely totally different efficiency.

To slim down the selection, corporations typically depend on LLM rating platforms, which collect person suggestions on mannequin interactions to rank the newest LLMs based mostly on how they carry out on sure duties.

However MIT researchers discovered {that a} handful of person interactions can skew the outcomes, main somebody to mistakenly imagine one LLM is the best alternative for a specific use case. Their research reveals that eradicating a tiny fraction of crowdsourced knowledge can change which fashions are top-ranked.

They developed a quick technique to check rating platforms and decide whether or not they’re vulnerable to this drawback. The analysis method identifies the person votes most chargeable for skewing the outcomes so customers can examine these influential votes.

The researchers say this work underscores the necessity for extra rigorous methods to judge mannequin rankings. Whereas they didn’t deal with mitigation on this research, they supply ideas which will enhance the robustness of those platforms, equivalent to gathering extra detailed suggestions to create the rankings.

The research additionally provides a phrase of warning to customers who could depend on rankings when making selections about LLMs that might have far-reaching and expensive impacts on a enterprise or group.

“We had been stunned that these rating platforms had been so delicate to this drawback. If it seems the top-ranked LLM will depend on solely two or three items of person suggestions out of tens of 1000’s, then one can’t assume the top-ranked LLM goes to be persistently outperforming all the opposite LLMs when it’s deployed,” says Tamara Broderick, an affiliate professor in MIT’s Division of Electrical Engineering and Laptop Science (EECS); a member of the Laboratory for Info and Determination Techniques (LIDS) and the Institute for Information, Techniques, and Society; an affiliate of the Laptop Science and Synthetic Intelligence Laboratory (CSAIL); and senior creator of this research.

She is joined on the paper by lead authors and EECS graduate college students Jenny Huang and Yunyi Shen in addition to Dennis Wei, a senior analysis scientist at IBM Analysis. The research might be offered on the Worldwide Convention on Studying Representations.

Dropping knowledge

Whereas there are numerous kinds of LLM rating platforms, the preferred variations ask customers to submit a question to 2 fashions and decide which LLM gives the higher response.

The platforms combination the outcomes of those matchups to supply rankings that present which LLM carried out finest on sure duties, equivalent to coding or visible understanding.

By selecting a top-performing LLM, a person possible expects that mannequin’s prime rating to generalize, which means it ought to outperform different fashions on their comparable, however not equivalent, software with a set of latest knowledge.

The MIT researchers beforehand studied generalization in areas like statistics and economics. That work revealed sure circumstances the place dropping a small share of information can change a mannequin’s outcomes, indicating that these research’ conclusions won’t maintain past their slim setting.

The researchers wished to see if the identical evaluation might be utilized to LLM rating platforms.

“On the finish of the day, a person desires to know whether or not they’re selecting one of the best LLM. If only some prompts are driving this rating, that means the rating won’t be the end-all-be-all,” Broderick says.

However it could be unattainable to check the data-dropping phenomenon manually. As an example, one rating they evaluated had greater than 57,000 votes. Testing an information drop of 0.1 % means eradicating every subset of 57 votes out of the 57,000, (there are greater than 10¹⁹⁴subsets), after which recalculating the rating.

As a substitute, the researchers developed an environment friendly approximation technique, based mostly on their prior work, and tailored it to suit LLM rating programs.

“Whereas we have now idea to show the approximation works beneath sure assumptions, the person doesn’t must belief that. Our technique tells the person the problematic knowledge factors on the finish, to allow them to simply drop these knowledge factors, re-run the evaluation, and verify to see in the event that they get a change within the rankings,” she says.

Surprisingly delicate

When the researchers utilized their method to widespread rating platforms, they had been stunned to see how few knowledge factors they wanted to drop to trigger vital modifications within the prime LLMs. In a single occasion, eradicating simply two votes out of greater than 57,000, which is 0.0035 %, modified which mannequin is top-ranked.

A special rating platform, which makes use of skilled annotators and better high quality prompts, was extra sturdy. Right here, eradicating 83 out of two,575 evaluations (about 3 %) flipped the highest fashions.

Their examination revealed that many influential votes could have been a results of person error. In some circumstances, it appeared there was a transparent reply as to which LLM carried out higher, however the person selected the opposite mannequin as an alternative, Broderick says.

“We will by no means know what was within the person’s thoughts at the moment, however possibly they mis-clicked or weren’t paying consideration, or they truthfully didn’t know which one was higher. The massive takeaway right here is that you just don’t need noise, person error, or some outlier figuring out which is the top-ranked LLM,” she provides.

The researchers recommend that gathering extra suggestions from customers, equivalent to confidence ranges in every vote, would supply richer info that might assist mitigate this drawback. Rating platforms might additionally use human mediators to evaluate crowdsourced responses.

For the researchers’ half, they wish to proceed exploring generalization in different contexts whereas additionally creating higher approximation strategies that may seize extra examples of non-robustness.

“Broderick and her college students’ work exhibits how one can get legitimate estimates of the affect of particular knowledge on downstream processes, regardless of the intractability of exhaustive calculations given the scale of contemporary machine-learning fashions and datasets,” says Jessica Hullman, the Ginni Rometty Professor of Laptop Science at Northwestern College, who was not concerned with this work. “The latest work gives a glimpse into the robust knowledge dependencies in routinely utilized — but additionally very fragile — strategies for aggregating human preferences and utilizing them to replace a mannequin. Seeing how few preferences might actually change the conduct of a fine-tuned mannequin might encourage extra considerate strategies for amassing these knowledge.”

This analysis is funded, partly, by the Workplace of Naval Analysis, the MIT-IBM Watson AI Lab, the Nationwide Science Basis, Amazon, and a CSAIL seed award.

Foreach, Spark 3.0 and Databricks Join

Artificial Intelligence

Dr. Mike
-

February 11, 2026
0

Behold the glory that’s sparklyr 1.2! On this launch, the next new hotnesses have emerged into highlight:

A registerDoSpark technique to create a foreach parallel backend powered by Spark that permits tons of of present R packages to run in Spark.

Assist for Databricks Join, permitting sparklyr to connect with distant Databricks clusters.

Improved assist for Spark buildings when accumulating and querying their nested attributes with dplyr.

Plenty of inter-op points noticed with sparklyr and Spark 3.0 preview have been additionally addressed just lately, in hope that by the point Spark 3.0 formally graces us with its presence, sparklyr might be absolutely able to work with it. Most notably, key options akin to spark_submit, sdf_bind_rows, and standalone connections at the moment are lastly working with Spark 3.0 preview.

To put in sparklyr 1.2 from CRAN run,

The total listing of modifications can be found within the sparklyr NEWS file.

Foreach

The foreach bundle supplies the %dopar% operator to iterate over components in a set in parallel. Utilizing sparklyr 1.2, now you can register Spark as a backend utilizing registerDoSpark() after which simply iterate over R objects utilizing Spark:

[1] 1.000000 1.414214 1.732051

Since many R packages are based mostly on foreach to carry out parallel computation, we are able to now make use of all these nice packages in Spark as nicely!

As an example, we are able to use parsnip and the tune bundle with knowledge from mlbench to carry out hyperparameter tuning in Spark with ease:

library(tune) library(parsnip) library(mlbench) knowledge(Ionosphere) svm_rbf(price = tune(), rbf_sigma = tune()) %>% set_mode("classification") %>% set_engine("kernlab") %>% tune_grid(Class ~ ., resamples = rsample::bootstraps(dplyr::choose(Ionosphere, -V2), occasions = 30), management = control_grid(verbose = FALSE))

# Bootstrap sampling # A tibble: 30 x 4 splits id .metrics .notes * 1 Bootstrap01 2 Bootstrap02 3 Bootstrap03 4 Bootstrap04 5 Bootstrap05 6 Bootstrap06 7 Bootstrap07 8 Bootstrap08 9 Bootstrap09 10 Bootstrap10 # … with 20 extra rows

The Spark connection was already registered, so the code ran in Spark with none extra modifications. We will confirm this was the case by navigating to the Spark internet interface:

Databricks Join

Databricks Join permits you to join your favourite IDE (like RStudio!) to a Spark Databricks cluster.

You’ll first have to put in the databricks-connect bundle as described in our README and begin a Databricks cluster, however as soon as that’s prepared, connecting to the distant cluster is as straightforward as operating:

sc <- spark_connect( technique = "databricks", spark_home = system2("databricks-connect", "get-spark-home", stdout = TRUE))

That’s about it, you at the moment are remotely linked to a Databricks cluster out of your native R session.

Buildings

Should you beforehand used acquire to deserialize structurally advanced Spark dataframes into their equivalents in R, you possible have observed Spark SQL struct columns have been solely mapped into JSON strings in R, which was non-ideal. You may additionally have run right into a a lot dreaded java.lang.IllegalArgumentException: Invalid sort listing error when utilizing dplyr to question nested attributes from any struct column of a Spark dataframe in sparklyr.

Sadly, usually occasions in real-world Spark use circumstances, knowledge describing entities comprising of sub-entities (e.g., a product catalog of all {hardware} parts of some computer systems) must be denormalized / formed in an object-oriented method within the type of Spark SQL structs to permit environment friendly learn queries. When sparklyr had the restrictions talked about above, customers usually needed to invent their very own workarounds when querying Spark struct columns, which defined why there was a mass standard demand for sparklyr to have higher assist for such use circumstances.

The excellent news is with sparklyr 1.2, these limitations not exist any extra when working operating with Spark 2.4 or above.

As a concrete instance, think about the next catalog of computer systems:

library(dplyr) computer systems <- tibble::tibble( id = seq(1, 2), attributes = listing( listing( processor = listing(freq = 2.4, num_cores = 256), value = 100 ), listing( processor = listing(freq = 1.6, num_cores = 512), value = 133 ) ) ) computer systems <- copy_to(sc, computer systems, overwrite = TRUE)

A typical dplyr use case involving computer systems could be the next:

As beforehand talked about, earlier than sparklyr 1.2, such question would fail with Error: java.lang.IllegalArgumentException: Invalid sort listing.

Whereas with sparklyr 1.2, the anticipated result’s returned within the following type:

# A tibble: 1 x 2 id attributes 1 1

the place high_freq_computers$attributes is what we might anticipate:

[[1]] [[1]]$value [1] 100 [[1]]$processor [[1]]$processor$freq [1] 2.4 [[1]]$processor$num_cores [1] 256

And Extra!

Final however not least, we heard about various ache factors sparklyr customers have run into, and have addressed lots of them on this launch as nicely. For instance:

Date sort in R is now appropriately serialized into Spark SQL date sort by copy_to

%>% print(n = 20) now truly prints 20 rows as anticipated as a substitute of 10

spark_connect(grasp = "native") will emit a extra informative error message if it’s failing as a result of the loopback interface isn’t up

… to only identify just a few. We wish to thank the open supply neighborhood for his or her steady suggestions on sparklyr, and are wanting ahead to incorporating extra of that suggestions to make sparklyr even higher sooner or later.

Lastly, in chronological order, we want to thank the next people for contributing to sparklyr 1.2: zero323, Andy Zhang, Yitao Li,
Javier Luraschi, Hossein Falaki, Lu Wang, Samuel Macedo and Jozef Hajnala. Nice job everybody!

If you want to atone for sparklyr, please go to sparklyr.ai, spark.rstudio.com, or among the earlier launch posts: sparklyr 1.1 and sparklyr 1.0.

Thanks for studying this publish.

Get pleasure from this weblog? Get notified of recent posts by e mail:

Posts additionally out there at r-bloggers

1...747576...395 Page 75 of 395